EPA-R2-72-007
August 1972 Environmental Protection Technology Series
Aerial Surveillance Spill
Prevention System
*i PRO^°
Office of Research and Monitoring
U.S. Environmental Protection Agency
Washington, D.C. 20460
-------
RESEARCH REPORTING SERIES
Research reports of the Office of Research and
Monitoring, Environmental Protection Agency, have
been grouped into five series. These five broad
categories were.established to facilitate further
development and application of environmental
technology. Elimination of traditional grouping
was consciously planned to foster technology
transfer and a maximum interface in related
fields. The five series are:
1. Environmental Health Effects Research
2. Environmental Protection Technology
3. Ecological Research
4. Environmental Monitoring
5. Socioeconomic Environmental Studies
This report has been assigned to the ENVIRONMENTAL
PROTECTION TECHNOLOGY series. This series
describes research performed to develop and
demonstrate instrumentation, equipment and
methodology to repair or prevent environmental
degradation from point and non-point sources of
pollution. This work provides the new or improved
technology required for the control and treatment
of pollution sources to meet environmental quality
standards.
-------
EPA-R2-72-007
August 1972
AERIAL SURVEILLANCE SPILL PREVENTION SYSTEM
By
C. L. Rudder
C. J. Reinheimer
J. L. Berrey
Contract No. 68-01-01^0
Project 15080 HOK
Project Officer
John Riley
Technology Division (ORM)
Environmental Protection Agency
Washington, B.C. 20^60
Prepared for
OFFICE OF RESEARCH AND MONITORING
U.S. ENVIRONMENTAL PROTECTION AGENCY
WASHINGTON, D.C. 20^60
For sale by the Superintendent of Documents, U.S. Government Printing Office
Washington, D.C. 20402 - Price $3.00
Stock Number 5501-00427
-------
EPA Review Notice
This report has been reviewed by the Environmental
Protection Agency and approved for publication.
Approval does not signify that the contents
necessarily reflect the views and policies of
the Environmental Protection Agency, nor does
mention of trade names or commercial products
constitute endorsement or recommendation for
use.
ii
-------
ABSTRACT
An aerial surveillance system, consisting of four Hasselblad cameras and
a Zeiss RMK 1523 camera, was evaluated for the remote detection of both
real and potential spills threatening inland waterways. Twenty-three
multiband and baseline missions were flown over oil refineries and other
industrial sites located adjacent to the Mississippi River. Baseline
flights were effective in counting storage tanks, locating and identify-
ing storage equipment and pipeline systems and determining dike
conditions. Stereoscopic analysis of baseline imagery was used to
estimate the height of tanks and dikes, drainage patterns and the area
of openly stored waste products. The multiband imagery was obtained by
combining each of nine filters with each of three different black-and-
white films. Spectral contrast image enhancement was accomplished by
either suppressing or transmitting the target reflected radiation
through proper film/filter selections. Spills, effluents and waste
areas were hence identified on the multiband imagery. Normal and false
color imagery was evaluated with the multiband imagery to determine the
best film/filter combinations for the areas of interest. Finally, the
personnel, equipment and procedures required to implement an aerial
surveillance spill prevention system were determined.
This report was submitted in fulfillment of Project //15080HOK, Contract
#68-01-0140, under the partial sponsorship of the Office of Research
and Monitoring, Environmental Protection Agency.
iii
-------
CONTENTS
Section Page
I Conclusions 1
II Recommendations 3
III Introduction 7
IV Design and Methods 9
V Experimental Program 15
VI Image Analysis 21
VII General Summary of Image Analysis 77
VIII Equipment, Personnel and Procedures for System
Implementation 81
IX Acknowledgements 85
X Appendices 87
-------
FIGURES
__PAGE_
1 PHOTOGRAPHS OF (A) ZEISS RMK 1523 CAMERA,
(B) HASSELBLAD CAMERA ARRAY, (C) AEROCOMMANDER,
AND (D) CESSNA 336. H
2 BASELINE IMAGERY OF OIL TANK STORAGE 22
3 ASPHALT BULK STORAGE AREA 24
4 GASOLINE BULK STORAGE AREA 26
5 BASELINE IMAGERY OF TITANIUM PLANT TANK STORAGE 27
6 BASELINE IMAGERY OF CHEMICAL PLANT TANK STORAGE 29
7 BASELINE IMAGERY OF STEEL PLANT WASTE LAGOON 31
8 STEEL MILL AND ADJACENT INDUSTRIAL WASTE DRAINAGE 32
9 BASELINE IMAGERY OF STEEL PLANT AND ADJACENT
INDUSTRIAL WASTE DRAINAGE 33
10 STEEL MILL AND ADJACENT INDUSTRIAL WASTE DRAINAGE 34
11 THERMAL INFRARED IMAGE OF STEEL MILL AND ADJACENT
INDUSTRIAL WASTE DRAINAGE 35
12 POWER PLANT FLY ASH POND 37
13 BASELINE IMAGERY OF WASTE OXIDATION PONDS 38
14 WASTE OXIDATION PONDS 39
15 WASTE OXIDATION PONDS 40
16 BASELINE IMAGERY OF OIL REFINERY EFFLUENT 41
17 OIL REFINERY EFFLUENT 43
18 BASELINE IMAGERY OF TITANIUM PLANT EFFLUENTS 44
19 TITANIUM PLANT EFFLUENTS 45
vi
-------
FIGURES (CONTINUED)
PAGE
20 THERMAL INFRARED IMAGE OF TITANIUM PLANT AND
EFFLUENTS 46
21 BASELINE IMAGERY OF SEWAGE PLANT EFFLUENT 47
22 SEWAGE PLANT EFFLUENT 48
23 BASELINE IMAGERY OF POWER PLANT EFFLUENT 49
24 GROUND TRUTH PHOTOGRAPH OF POWER PLANT EFFLUENT 50
25 THERMAL INFRARED IMAGE OF POWER PLANT EFFLUENT 51
26 BASELINE IMAGERY OF OIL REFINERY WASTE STORAGE 52
27 LIME SLUDGE WASTE AREA 53
28 BASELINE IMAGERY OF OIL REFINERY WASTE AREA 55
29 OIL WASTE AREA 57
30 BASELINE IMAGERY OF OPEN SULFUR STORAGE 59
31 SULFUR STORAGE AREA 59
32 BASELINE IMAGERY OF OPEN STORAGE OF COAL 60
33 COAL STORAGE AND LOADING AREA 61
34 BASELINE IMAGERY OF OPEN SHALE QUARRY 62
35 SHALE QUARRY 63
36 (A) BASELINE AND (B) GROUND TRUTH IMAGERY OF
BARGE LOADING AREA 65
37 (A) BASELINE AND (B) GROUND TRUTH IMAGERY OF
BARGE LOADING AREA 66
vii
-------
FIGURES (CONTINUED)
PAGE
38 BASELINE IMAGERY OF CHEMICAL PLANT ACID LOADING
AREA 68
39 SULFURIC ACID LOADING AREA 69
40 BASELINE IMAGERY OF OIL REFINERY LOADING AREA 70
41 BASELINE IMAGERY OF OIL REFINERY 72
42 THERMAL INFRARED IMAGE OF OIL REFINERY 74
43 OBLIQUE IMAGERY OF OIL REFINERY 75
44 IMAGERY RECORDED ON A (A) HAZY AND (B) CLEAR DAY 76
45 IMAGERY OF SAME AREA PHOTOGRAPHED ON SEQUENTIAL
DAYS 76
viii
-------
TABLES
No. Page
1 Filter Factors 12
2 Flight Program 16,17
3 Simultaneous Ground Truth Data 18,19
IX
-------
SECTION I
CONCLUSIONS
The feasibility of an aerial surveillance spill prevention system for
identifying and locating actual or threatened discharges of oil and
hazardous polluting substances from onshore facilities into or upon
inland waterways has been demonstrated. In addition, the personnel,
equipment and procedures required to implement an aerial spill preven-
tion system have been determined.
Baseline photographs, at a scale of approximately 1:4000, are useful for
identifying potential spill threats to inland waterways from storage
equipment, storage and processing tanks, pipeline systems, dike condi-
tions and the presence of trash or debris in diked areas. Stereoscopic
analysis of this imagery also permits estimating dike heights, drainage
patterns and the runoff patterns from openly stored raw materials, all
of which are aids in identifying potential spill threats.
Positive identification of spills as oil or hazardous materials is
necessary to the determination of the potential threat. Multiband
photographs are effective in providing spectral contrast enhancement for
the identification of oil and oil waste in oil refinery areas. Multi-
band imagery at scales of 1:9000 or larger was found effective for
detecting spills, effluents, and waste areas.
For aerial multiband photography, the preferred Kodak film/filter
combinations for the detection of oil, oil derivatives and oil waste
products are 2403/32, 2403/99, 2424/32 and 2424/99. Combinations
2424/99 and 2424/32 are recommended for detecting oil and oil deriva-
tives spilled on soil. When the background is water, 2409/99 and 2403/32
are recommended. With both film types, filter 99 provides maximum image
contrast of oil while filter 32 best distinguishes oil drainage patterns.
Of the color films investigated, Kodak type 2448 permits the most
effective detection of spills, effluents, and raw materials at industrial
sites. The color photographs provide cues for the identification of many
hazardous materials.
From available thermal infrared (8,000 to 14,000 nm) imagery of the same
sites, it was determined that processing facilities, materials, and
effluents which are "warm" can be readily detected. Such information
provides useful indicators of activity. However, this additional infor-
mation is not considered to be essential.
Aerial surveillance for the detection, identification, and location of
actual and threatened discharges of oil and hazardous materials into
waterways can be accomplished at reasonable cost with a photographic
system. The system should be capable of producing both color and multi-
band photographs. When properly analyzed, such photographs will provide
sufficient information for the assessment of real and potential threats.
-------
SECTION II
RECOMMENDATIONS
The following recommendations are made for the implementation of an
aerial surveillance spill prevention system for monitoring real and
potential threats to inland waterways.
Aircraft
The selected aircraft should have good flying qualities.at altitudes of
1500 ft and above at air speeds above 100 knots. Its power plant
should be adequate to maintain ground speeds of 120 knots during photo-
graphic operations. Interior space should be adequate for a crew of
three and for installation and efficient access to the photographic
subsystem described below.
Photographic Subsystem
For multiband photography an array of four 70-mm cameras, with film
magazines to accommodate at least 15 ft of film, is recommended.
Motorized film advance and electrical operation of camera shutters is
essential. Shutters should permit exposure of 1/250 to 1/500 second to
reduce image smear to an acceptable level for ground speeds of 110-120
knots. The cameras in the array should be mounted with optical axes
vertical. The quad-array should be attached to a mounting plate which
is compatible with the aircraft camera mount. The combination provides
adjustment for drift angle and leveling in two axes. The focal length
of the camera should provide a minimum image scale of 1:9000 at operating
altitudes. A baseline camera with 9-in. by 9-in. format, is recommended
for baseline and stereographic photography. For best results, the map-
ping camera axis must be vertical in flight. The focal length of the
camera should provide an image scale of 1:4000 at the operating alti-
tudes. A view finder is recommended for determining start and stop
times for the cameras. The camera control should include an interval-
ometer to provide properly spaced commands to insure correct overlap
in successive frames.
Sixty percent overlap should be maintained in stereophotography. A
command unit is required to initiate simultaneous exposures on the
multiband camera array. Filters, filter holders, film cassettes, and
other loose parts complete the photographic subsystem.
Flight Crew
A crew of three, pilot, aerial photographer, and camera monitor, is
recommended. The photographer's primary tasks are to select the initial
and final camera exposures, monitor overlap and maintain camera level
and crab. The camera monitor's tasks are to watch for possible camera
malfunctions, and to monitor air traffic in the operating areas.
-------
Film Processing
Precision film processing is required for both the multiband and base-
line photography. The Versamat continuous roll film processor or
equivalent is recommended over reel and tank developing. The Versamat
represents a higher initial cost, but is more efficient for long term
usage, provides quick image processing and requires minimum additional
dark room accessories. If the aerial film is processed at the user's
facility, a photoprocessing technician will be required to operate and
maintain the photoprocessing equipment. Many aerial photographers are
also capable of performing these duties.
Image Analysis
For an image interpretation station, a standard light table capable of
handling a 9-in. film format, a stereoscopic viewer, 10X and 30X
magnifiers and a parallax bar are the minimum recommended equipment.
The number of image interpretation stations and the total personnel re-
quirement will be determined by the volume of imagery to be analyzed,
the scale and quality of the aerial photographs, the requirement for
detailed stereographic analysis, and the complexity of required reporting.
Ground Truth
As an adjunct to the aerial surveillance system, a capability for deter-
mining ground truth is recommended. Visits to sites under aerial sur-
veillance for ground level observation of spill threats, examination of
protective measures, and collection of samples of suspect pollutants for
on-site or later analysis can provide validation of the results of photo-
graphic analysis. Flight crew personnel and photointerpreters should be
given the opportunity to participate in ground truth operations.
Film and Filters
Kodak Film types 2405 and 2448 are recommended for taking black-and-white
or color 9-in. format baseline photography, respectively. Similarly,
Kodak Film type 2403 is recommended for taking the multiband photography
and film type 2448 is recommended for taking color photography when
required in the Hasselblad camera array. Kodak filters 99 and 32 are the
minimum filters recommended to obtain the multiband photographs over oil
refinery areas. Color, rather than multiband photography is presently
recommended for monitoring industries using or manufacturing hazardous
materials. Because the 70-mm cassettes are not light tight, camera
magazine loading should be done in a dark room facility.
Infrared Thermal Imagery
An infrared thermal imaging system is not recommended as part of the
aerial surveillance spill prevention systems as sufficient information
for adequate monitoring can be obtained from a multiband and baseline
photographic system.
-------
Weather Conditions
The multiband and baseline photographs should be taken on clear days
if possible. On hazy days, adequate multiband and baseline images can
be recorded if the appropriate adjustment of exposure is made. Additional
information on the effects of adverse weather conditions on multiband
imagery is needed to extend the effectiveness of an aerial surveillance
spill prevention system.
-------
SECTION III
INTRODUCTION
The purpose of the project was to demonstrate the feasibility of an
aerial surveillance system for the detection and location of real and
potential spill threats to the inland waterways. It was initially postu-
lated that conventional and multiband aerial photographs would provide
sufficient data to accomplish this task. Results, reported herein, demon-
strated that an aerial surveillance system can provide timely information
at reasonable cost, for the detection, identification and threat assess-
ment of real and potential spills of oil, oil derivatives and waste and
other hazardous materials. In addition, preferred film/filter combinations
were identified for the detection and identification of spilled materials
using multiband photography.
The St. Louis area has a variety of industrial complexes that are located
near or adjacent to the Mississippi River. A survey was made of imagery
in the McDonnell Reconnaissance Laboratory Data Base and site selection
was made with the guidance of the Environmental Protection Agency. These
industries include two oil refineries, two barge loading facilities, a
titanium plant, a chemical plant, a cement manufacturing plant, a sewage
disposal plant, a steel mill and a thermo-electric plant. Site selection
was based on proximity to the Mississippi River and the presence of
industrial effluents and potential spill threats. Some of the industries
are located behind a levee or flood wall which prevents direct surface
drainage from these sites to the river. These non-adjacent river loca-
tions were investigated for two reasons. First, these sites discharge
waste materials through public sewers or private drainage systems and
therefore covertly threaten the inland waterway. Second, the analysis
applied to these areas demonstrates the system capabilities that can be
applied to other industries located adjacent to inland waterways.
Location and identification of real and potential spill threats requires
determination of drainage patterns, dike heights, tank size, and the
volume of detention ponds and waste areas. Such dimensional data can be
derived from aerial photographs obtained with cameras capable of providing
high geometrical fidelity and stereoscopic coverage. Consequently, the
aerial surveillance system included a high quality mapping camera. Base-
line flights used the mapping camera to record, in color and in black-and-
white, imagery of the selected sites for detailed analysis.
Multiband photography combines various spectral filters with films having
different spectral responses and provides contrast image enhancement of
materials having particular reflective characteristics. By a judicious
choice of film and filters, material identification can be made through
the use of multiband photography. Positive identification of spilled
material as oil or hazardous material is necessary to the meaningful
assessment of a real or potential threat to inland or coastal waterways.
-------
The aerial system employed an array of four identical cameras for simul-
taneous recording of site images. Each camera was equipped with a dif-
ferent film/filter combination to produce the multiband photographs.
Ground truth is the term applied to a body of collected data gathered
at the industrial sites of interest to aid and confirm the analysis of
aerial imagery. On their initial plant visit, ground truth teams used
mosaics constructed from baseline photographs to examine potential
threat areas that included processing, storage and transportation facil-
ities. Each area was labeled on the mosaic and pertinent data on waste
treatment and removal techniques was kept in an appropriately labeled
log book. Similarly, ground truth photographs of each area were taken
where possible, with a 35-mm hand held Nikon camera using Etachrome X
film and appropriately labeled. From the initial ground truth informa-
tion, the multiband flight areas were selected. The ground truth photo-
graphs and descriptions of specific areas were used to estimate the
spectral signatures of the area of interest. Ground truth data relevant
to aerial photography was also taken simultaneously with one multiband
flight at each industrial site.
Advance arrangements were made with each company visited to allow the
entry of ground truth teams. Initially, each company was notified by
phone of the ensuing project and meetings were arranged to explain the
program in depth. At these meetings, the evaluation of the aerial sur-
veillance system rather than the monitoring of individual companies, was
emphasized as the objective of this project. In general, sincere interest
was expressed in the potential use of an aerial surveillance spill preven-
tion system and a high degree of cooperation was obtained. Permission to
take ground truth photographs was granted by most companies except for
areas where government or company regulations prohibited photography. For
this cooperation, each company was given a chance to review the draft report
and was promised a copy of the final report. From knowledgeable repre-
sentatives at each site, the ground truth team gained valuable understand-
ing of plant functions, procedures and resources not readily apparent in
the aerial photographs.
As with any photographic system dependent upon natural lighting, success-
ful application requires adequate sunlight and a reasonably transparent
atmosphere. No flights were attempted on overcast days, but useful results
were obtained through light haze. Although the effects of air turbulence
were not specifically evaluated, it was observed that gusty air can degrade
both photographic image quality and air crew performance.
The following sections of this report describe in detail the aerial system
used and the findings of the experimental program.
-------
SECTION IV
DESIGN AND METHODS
Camera System
The baseline photographic system requires a conventional mapping camera
capable of producing images of high geometrical fidelity. This type of
camera employs a six inch focal length lens and supplies a nine inch
square film format. The camera system must fit an aircraft mount that
provides leveling in two axes and drift compensation. A variety of cameras
satisfy these requirements: the Zeiss RMK 1523 and RMK-A-15/23, the
Fairchild CA-8 and T-ll, the Wild Heerbrugg RC-8 and the Carl Zeiss Jena
11.5/188. Since it was readily available, the Zeiss RMK 1523 mapping camera
shown in Pig. la was used in this project. A Zeiss intervalometer was
used to provide the shutter trip pulses and adjustment of overlap on the
baseline imagery. The associated Zeiss view finder allowed the aerial
photographer to select the initial and final exposure points.
The multiband camera system consists of an array of four 70-mm cameras,
an aircraft camera mount, a camera exposure command unit, and a view
finder. The cameras must have a minimum film capacity of 15 ft, have
automatic film advance and shutter cocking, and at least a 2 to 3 in. focal
length lens. In addition, the dimensions of the cameras should be such
that four cameras could be mounted in the aircraft camera window. Poten-
tial multiband cameras include the Fairchild type CAX-12, the Aerojet
Delft TA-7M, the Hasselblad 500 EL/M, the Itek KA-61, and the Naval Air
Development Center X70-7. The Hasselblad 500 EL/M camera with a 50 mm,
f/1.4 Distagon lens satisfied all the requirements for the minimum price
and was chosen for this project.
The cameras were attached to a rigid plate which was compatible with
the aircraft mounting plate located in the camera window. The latter
was adapted from the Zeiss camera assembly. The combination provided
drift adjustment and leveling in two axes. The leveling was accomplished
with a spirit level attached to the rigid plate. The camera mounting
plate was fabricated from 0.25-in. aluminum plate. Four cameras were
held by quick release mounts with the cameras oriented perpendicular
to the plate, as shown in Fig. Ib. The camera command unit provided
a relay for simultaneously activating all four cameras. The Zeiss
aerial mapping view finder was used with the Hasselblad array.
The camera magazines hold the metal cassette that takes 15 ft of film
which can provide seventy exposures. These cassettes are only for holding
the film and have very poor light seals. This makes magazine loading in
dark room conditions necessary. For inflight film loading in the cameras,
a second set of magazines is required.
-------
Aircraft Requirements
The camera system dictates the aircraft requirements. A minimum image
scale of 1:5000 was considered satisfactory for performing mensuration
on the baseline photographs. For the mapping camera's 6 inch focal
length lens, the image scale would be achieved for aircraft altitudes
up to 2500 ft above ground level (AGL). Similarly, the scale for the
Hasselblad camera photographs calculates to 1:9000 for an altitude of
1500 ft AGL. For the multiband analysis, this scale was considered
minimal. At these altitudes, and for camera exposure times of 1/250
seconds, the aircraft ground speed should be approximately 110 to 120
knots to reduce image smear to an acceptable level. These ground
speeds were also commensurate with the camera cycling rates. It should
be pointed out that longer focal length lenses would allow higher
altitudes and faster aircraft speeds. More sophisticated camera systems
introduce an increase in cost.
For the aforementioned camera system, the aircraft must be capable of
stable flight at 1500 ft AGL, and at ground speed of 110 to 120 knots.
Camera accessibility during flight is needed for monitoring camera per-
formance. The Aerocommander Model 680 and Cessna 336 aircraft used
for this project are shown in Figs. Ic and Id respectively. The Aero-
commander satisfied all the aforementioned requirements, had a ceiling
of 20,000 ft and a maximum flight time of 5-1/2 hours. The Cessna 336
satisfied all requirements with the exception of camera accessibility
during flight. The Cessna's maximum ceiling was 10,000 ft and its
maximum flight time was 5-1/2 hours.
Flight Crew
The flight crew consists of the pilot, the aerial photographer and the
camera monitor. The aircraft pilot should be qualified in aerial survey
flight operations, including proficiency in the use of large scale maps
(e.g., 1:24,000). He must be able to maintain the flight line to within
- 3 deg and must be able to maintain the aircraft attitude while holding
the flight line. The aerial photographer must have experience in aerial
photography and be able to lay out the desired flight plan. Because of
the possibility of camera failure, a third crew member should accompany
each flight to monitor camera performance. If a camera malfunction does
occur, the photographic run can be terminated, the malfunction remedied
if possible, and the photographic run continued. This procedure re-
quires more manhours per flight hour, but reduces total flight cost by
improving the number of successful missions. Besides monitoring the
camera performance, the camera monitor has an additional duty to look
for and keep track of any other aircraft in the area. During a photo-
graphic run, the pilot is entirely occupied with the flight line and
aircraft altitude and the photographer is totally occupied over the
view finder. The camera monitor can continually maintain a lookout for
other aircraft. This is especially important over a metropolitan area
where the air traffic is heavy.
10
-------
(a)
(b)
(d)
Figure 1 Photographs of (a) Zeiss RMK 1523 camera, (b) Hasseblad
camera array, (c) Aerocommander and (d) Cessna 336
Filter Factors
Whenever a filter is placed in front of a camera system, the exposure
(f/stop and exposure time) needed to achieve the same density as obtained
without the filter can change drastically. The amount of this change
depends primarily on the filter spectral transmittance, the film spectral
sensitivity, and sunlight spectral characteristics. Over the period of
this project, the sun's spectral characteristics do not change suffi-
ciently to affect the filter factor. (From 27 July 1971 to 22 November
1971 the sun-zenith angle at local apparent noon varied from 17 to 50
deg.) Consequently, the solar energy must propagate through different
atmospheric masses, so that the solar spectrum is affected in two ways.
First, the total solar power per wavelength changes by approximately a
factor of 3 over the specified time period. Second, there is an approx-
imate 20% spectral power shift between 450 and 850 nm. Since the
! :
-------
film's latitude allows satisfactory images to be recorded of objects
whose incident power is a factor of 8 on either side of the best exposure
neither of these effects will significantly affect the filter factor. '
After film types 2403, 2424, and 2475 were loaded in separate Hasselblad
cameras, photographs of a gray scale and a color chart were taken
outside at various f/stops, both with and without the nine filters
used in this flight program. By comparing "normal" photographs to the
filtered photographs taken at various f/stops, the f/stop correction for
each film/filter combination was determined. Table 1 lists the f/stop
correction (or filter factor) for each film/filter combination.
Table 1 Filter factors
Filters
18A
39
47B
32
35
65
75
98
99
Film types
2403
3
1
2
1/2
1/2
2
4
2
3
2475
3
1
2
1/2
1/2
2
4
2
3
2424
2-1/2
1/2
1
1/2
1
1-1/2
1
1/2
1
Because the filter f/stop correction factors were determined on the
ground rather than from aerial photographs, sometimes an additional
1/2 stop was needed to achieve the best image. The latitude of the
film, however, allows satisfactory images to be obtained within 1 to 2
stops on either side of the best f/stop for similar atmospheric con-
ditions. During hazy days, additional compensation of the exposure is
needed and can be computed from an Aerial Index Exposure Calculator.
Since almost all flights were made on clear or hazy days, complete
experimental data is not available on the correction factors needed
for other atmospheric conditions.
12
-------
Flight Parameters
Mission planning for the acquisition of aerial imagery is required for
each new area to be photographed. A well-designed flight plan ensures
the acquisition of a minimum amount of imagery which will be used to
produce the required information concerning a specific area. The
following factors were considered to define the flight plans used to
acquire imagery for this project.
Film/Filter Combinations - The type of film and filter to be used is
determined by the target and the purpose of the flight. In the base-
line flights, a high resolution, black-and-white, stable base film and
appropriate haze filter were required for mensuration work. In the
multiband photography, the spectral signature of the target determined
the most advantageous film and filter.
Imagery Scale Factor - The scale factor of a photograph is the ratio
of the photo distance to the ground distance i.e., one inch on the photo-
graph equals twelve hundred inches on the ground means a scale of 1:1200.
For vertical photographs, scale is computed by dividing the aircraft
altitude above ground by the focal length of the camera lens. This,
along with the resolution capability of the film and camera, will deter-
mine the smallest object on the ground which can be recorded at the film
plane. In this project it was decided that mensuration was to be
performed using baseline imagery (high resolution) and tonal information
was to be acquired using the multiband photography (less resolution).
Exposure Determination - The proper film exposure is computed by using
the sun angle (latitude of target, time of year, and time of day),
altitude (above ground level), film sensitivity (Aerial Exposure Index
or Aerial Film Speed) , shutter speed (in fractions of a second), ground
haze conditions and film/filter combination. Shutter speed is then
compared to aircraft ground speed to determine if the image smear
(movement of image on film plane during exposure) is within allowable
limits.
Flight Line Plan - The target is located and outlined on a map of the
appropriate scale, typically 1:24,000 for an industrial area. The
ground area within the field of view (FOV) of the camera for the selected
flight altitude is plotted to scale and superimposed on the target area.
If the width of the target is less than the frame coverage, only one
flight line is required for that target. The center of the first frame
is plotted on the map at the edge of the target area and all additional
frame centers are plotted till the target length is covered. For
stereo coverage of the target, the photographs should overlap by 60%.
For monoscopic coverage, the overlap should be only 15% to assure
continuous coverage.
If the target area is wider than the photograph, additional flight lines
are plotted. The flight line overlap should be from 15 to 25%, depending
13
-------
on terrain elevation variation, wind conditions, and the pilot's abil-n-
to fly the indicated flight line. ability
Each target area requires a separate plan for proper coverage with a
minimum of photographs. The flight lines with the beginning and end <=
the photographic run are plotted in a distinct color for ealy^eaS*
One copy is prepared for the pilot and one copy for the photographed*
Flight Plan Sheet - The flight plan sheet is a summary of the
—>-., i-x.-i-siii. Lj-me, cne names of
the crew, date, number of exposures per flight line, heading, and time
of completion of each flight line.
Film and Filter Selection
Kodak Tri-X Aerographic film type 2403, Kodak Infrared Aerographic film
type 2424 and Kodak Recording film type 2475 were used and evaluated
in this project. Only the film type number rather than the descriptive
title will be specified in this report. Kodak filters 18A, 47B, 39,
32, 35, 65, 75, 98, 99 and 25 were combined with the above mentioned
films to determine the best film and filter combination for multiband
identification of oil and hazardous material spills. The various film
and filter combinations are specified film/filter. Hence the use of
film type 2403 and filter 32 will be denoted 2403/32. The film and
filter spectral characteristics along with their uses in multiband
photography are discussed in Appendix A.
Photographic Processing System
Precision photographic processing of the 70-mm multiband film and color
film is necessary for the extraction of scientific data from the
resultant images. Precision processing insures that the tonal variation
in the multiband images is a result of the various targets' spectral
reflectivities and not the result of fluctuation of the processing
techniques. The essential parameters and their influence in deter-
mining the final precision processing system are discussed in detail in
Appendix B.
14
-------
SECTION V
EXPERIMENTAL PROGRAM
From 27 July 1971 to 22 November 1971, twenty-three flights were made
over the St. Louis area. Three of these were baseline flights, while
the remainder were multiband flights. The multiband flights were
divided into two general areas: the oil refineries and the remaining
industrial sites. This division was necessitated by the limited
number of exposures (70) on each flight. The oil refinery industry
was divided into eight aerial flight lines. The remaining six
industrial areas were divided into nine aerial flight lines. Table 2
lists these flights and includes the general area, the flight date,
the film/filter combinations on each camera, the filter corrected
f/stops, exposure time, altitude, and overall performance. The last
parameter only lists gross failures such as camera malfunctions and
film fogging. Detailed image analysis is not included here.
Simultaneous ground truth measurements were made at least once at each
industrial site while multiband missions were being flown. The param-
eters of interest were wet bulb and dry bulb temperatures, ground
temperature, barometric pressure, sun angle, wind velocity and
direction, and luminosity measurements in each of the spectral bands
characterized by the filters flown on that mission. These measurements
were made at two locations at each site. Table 3 lists these param-
eters recorded during the multiband missions and includes the site
location, time, and date. Photographs of each location were taken on
35-ram Tri-X film with a Nikon camera with the same filters being flown.
General correlation of the aerial multiband and ground imagery was
observed.
15
-------
Table 2 Flight program
Date
27Jul71
30Jul71*
14 Sep 71
28 Sep 71
60ct71
11 Oct71
120ct71
14Oct71
15 Oct 71
28 Oct 71
28 Oct 71
1 IMov 71
2 Nov 71
3 Nov 71
3 Nov 71
4 Nov 71
4 Nov 71
5 Nov 71
10 Nov 71
11 Nov 71
11 Nov 71
12 Nov 71
22 Nov 71
Altitude (ft)
(above ground
2300
1300
2500
2000
1200
1200
1200
1200
1500
1500
1500
1500
1500
1500
1500
1500 to
2000
1500 to
2000
1500 to
2000
1500 to
2000
1500 to
2000
1500 to
2000
1500 to
2000
1500 to
2000
1500 to
2000
1500 to
2000
1500 to
2000
3000
Flight
area
Oil refinery
Steel mill
Other industries
Oil refinery
Other industries
Oil refinery
Oil refinery
Oil refinery
Oil refinery
Oil refinery
Oil refinery
Oil refinery
Oil refinery
Oil refinery
Oil refinery
Oil refinery
Other industries
Other industries
Other industries
Other industries
Other industries
Other industries
Other industries
Dther industries
Other industries
Other industries
Oil refinery
Other industries
Camera 1
Film type
2405
2448
2403
2475
2403
2475
2403
2403
2403
2475
2424
2424
2424
2403
2403
2403
2475
2475
2424
2424
2424
2424
2424
2475
2405
Filter
#12
KLF
47B
47B
35
35
18A
65
75
32
18A
32
75
47B
35
75
47B
35
18A
32
75
32
18A
18A
#12
t#
8
5.6
b.ti
8
4
5.6
5.6
8
4
22
6.3
13
9.7
8
16
4
8
16
6.3
13
11
13
6.3
5.6
8
Speed
1/550
1/275
1/500
1/500
1/500
1/500
1/500
1/500
1/500
1/500
1/500
1/500
1/500
1/500
1/500
1/500
1/500
1/500
1/500
/500
/500
/500
/500
/500
/550
Camera 2
Film type
2403
2475
2403
2475
2403
2403
2403
2475
2424
2424
2424
2403
2403
2403
2475
2475
2424
2424
2424
2424
2424
2475
hlter
39
39
32
32
39
47B
98
35
39
35
98
39
32
98
39
32
39
35
98
35
39
39
f#
8
11
11
16
11
8
8
16
— •
13
11
9.7
11
22
8
11
22
13
11
13
13
13
11
Speed
1/500
1/500
1/500
1/500
1/500
1/500
1/500
1/500
1/500
1/500
1/500
1/500
1/500
1/500
1/500
1/500
1/500
1/500
1/500
1/500
1/500
1/500
• Base flight - Zeiss camera
16
-------
Table 2 Flight program (cont.)
Camera 3
Film type
2403
2475
2403
2475
2403
2403
2403
2443
2424
2424
2424
2403
2403
2403
2475
2443
2424
2424
2424
2424
2424
2475
Filter
ISA
ISA
65
65
47B
32 + .6
neu. den.
99
12
CC10M
47B
65
99
ISA
65
99
ISA
12
CC10M
47B
65
99
65
47B
47B
f#
4
5.6
5.6
8
8
16
5.6
5.6
11
9.7
8
5.6
8
5.6
5.6
5.6
11
9.7
11
9.7
11
8
Speed
1/500
1/500
1/500
1/500
1/500
1/500
1/500
1/500
1/500
1/500
1/500
1/500
1/500
1/500
1/500
1/250
1/500
1/500
1/500
1/500
1/500
1/500
Camera 4
Film type
2448
2448
2448
2448
2448
2448
2448
2403
2424
2448
2403
2448
2448
2403
2448
2448
2424
2424
2448
2443
2403
2424
Filter
—
— "
-
HF3
HF3
32
25
HF3
35
HF3
HF3
47B
HF3
HF3
25
99
HF3
12
CC10M
47B
39
f#
4
4
4
4
4
5.6
5.6
22
11
5.6
11
5.6
5.6
8
5.6
5.6
11
8
5.6
5.6
8
13
Speed
1/250
1/250
1/250
1/250
1/250
1/250
1/250
1/500
1/500
1/250
1/500
1/250
1/250
1/500
1/250
1/250
1/500
1/500
1/250
1/250
1/500
1/500
Performance
Good imagery
Good imagery
2403 overexposed -
camera #4 malfunctioned
2403 overexposed
cameras #1, #2 malfunctioned
Camera 3 out of focus
and jammed
Camera #3 1 stop underexposed
Good imagery
Good imagery
Camera #2 malfunctioned
2424 film fogged on some frames
Some fogging
Camera #1 malfunction
Good imagery
Good imagery
Camera #2 jammed
fixed in flight
Good imagery
All lost because of
film fogging
All lost because of
film fogging
Good imagery
Good imagery
Good imagery
Good imagery
Good imagery
GP71-1642-58
17
-------
Table 3 Simultaneous ground truth data
Industry
Oil refinery
Oil refinery
Cement plant
Chemical plant
Titanium plant
Power plant
Sewage plant
Steel mill
Date
14 Sep 71
28 Sep 71
1 IMov 71
2 Nov 71
2 Nov 71
3 Nov 71
4 Nov 71
5 Nov 71
Location
Oxidation
ponds
Bulk storage
area
Loading area
Oil waste ponds
Shale quarry
Effluent
Water tower
Benzene tank
Railroad track
adjacent river
Storage area
Effluent
Fly ash pond
Effluent
Large pond
Final pond
Time
11:40 a.m.
12:10 p.m.
1:25 p.m.
2:00 p.m.
1:10 p.m.
1:45 p.m.
11:05a.m.
11:16 a.m.
1:12 p.m.
1:30 p.m.
11:20 a.m.
11:45 a.m.
11:20 a.m.
9:45 a.m.
10:30 a.m.
Wet bulb temp
°F
72.5
70.0
77.0
78.0
68.0
68.0
57.0
56.0
56.5
56.0
50.0
50.0
52.0
52.0
49.0
Dry bulb tamp.
°F
87.0
86.0
87.0
92.5
77.0
78.0
67.0
67.0
70.0
70.0
61.0
59.5
62.0
64.0
64.0
Ground temp.
°F
85.0
95.0
93.0
87.0
72.0
73.5
65.5
66.5
83.0
54.0
66.0
60.0
68.0
59.0
64.0
GP71-1642-77
18
-------
Table 3 Simultaneous ground truth data (cont.)
Sun
angle
(deg)
51.0
54.0
48.0
49.0
33.0
30.0
37.0
36.5
33.0
31.5
35.0
36.0
36.0
21.0
25.0
Wind velocity
and direction
mph
0
2.0
4-10
10-12
6
2
2-4
11
0-4
0-2
10-12
20-30
3-6
12
8-10
direction
sw
s
s
ssw
ssw
w
w
sw
sw
sw
sw
SSE
ESE
ESE
Barometric
pressure (in.)
30.02
30.04
30.16
30.13
30.04
30.03
31.70
31.70
31.20
31.40
30.34
30.30
30.32
31.90
32.00
Spectral luminosity
Filter
None
47B
39
ISA
None
47B
39
18A
None
35
32
65
None
35
32
65
None
18A
39
47B
None
47B
39
ISA
None
35
32
65
None
32
35
65
None
32
35
65
None
32
35
65
None
18A
47B
39
75
98
99
None
ISA
39
47B
75
98
99
None
ISA
39
478
32
35
None
32
35
65
99
None
32
35
65
99
Foot candles
6000
390
875
260
6000
390
875
260
6000
1500
3500
500
6000
1500
3500
500
7000
390
1000
390
2000
113
260
113
6000
1500
3500
500
6000
3500
1500
500
6000
3500
1250
500
6000
3500
1250
500
6000
260
390
875
390
390
750
6000
260
875
390
390
325
750
7200
312
1000
390
4000
1500
6000
3250
1500
625
750
6000
3500
1500
500
750
Weather
conditions
Slight haze
Slight haze
Clear
Clear
Partially cloudy
Cloudy
Clear
Clear
Clear
Clear
Clear
Clear
Clear
Clear
Scattered
clouds
Clear
Scattered
clouds
19
-------
SECTION VI
IMAGE ANALYSIS
Technique
From the 17 aerial flight lines, 20 individual areas covered with one
to three 70-mm frames were chosen for detailed multiband image analysis.
The best image or images were chosen for each film type.
The best film/filter combinations were then chosen by comparing the
best filtered images recorded. These images were grouped into general
categories, such as bulk storage areas, effluents, waste lagoons, etc.,
and resulting best film/filter combinations were compared and evaluated.
The multiband image evaluations contained in this report are based on
an examination of the original transparencies and not on the prints con-
tained herein. Baseline photographs were chosen of the same categorical
areas for detailed mensuration work.
The information obtainable from the baseline photographs, the value of
multiband imagery and the best film/filter combinations, and the value
of thermal infrared imagery extracted from our data base are discussed
for each area.
The imagery recorded on film types 2475 and 2443 is not mentioned in
the multiband imagery discussion. The 2475 imagery was found to have
less contrast than the 2403 imagery and was eliminated as a multiband
imagery film. The slight increased ultraviolet sensitivity of film
type 2475 was offset by its lack of contrast. The imagery recorded on
film type 2443 was found to be unique in its color scheme and
emphasizes vegetation detection. Straight color imagery obtained with
film type 2448 offers as much in contrast and detection capabilities as
film type 2443 for the areas under consideration.
Oil Tank Storage
The baseline imagery shown in Fig. 2 is typical of a large oil tank
storage area and shows many of the features of this type of installation.
While this area is located 21/2 miles east of a major river, it does
not constitute a real threat. The value of examining this major storage
area is the application of this analysis to similar areas located adja-
cent to an inland waterway which could constitute a real or potential
threat.
The overall slope of the ground in Fig. 2 is to the northeast or lower
left corner of the photograph. The drainage is to a swampy area (not
shown in the photograph) with no apparent outlet. At the center of
the image, labeled A, there is a complex of waste ponds, with roads for
21
-------
Figure 2 Baseline imagery of oil tank storage
dumping and removing waste material. The area has been used to dispose
of waste containing considerable solid material. The dikes are approxi-
mately 5 ft high, and, in some cases, the precipitated solid matter has
completely filled the ponds, and would constitute a potential threat if
located adjacent to an inland waterway. At point B is a square-shaped
area where the dike is only approximately 3 ft high. This area does
not represent a threat as it only encloses a flare which is part of the
safety system used for the burning of gasses during plant emergencies.
The row of tanks, marked C, are of the floating roof variety used to
store gasoline. These tanks have a diameter of 128 ft and a height of
43 ft to the upper rim. They are easily identified by the walkway
extending from the rim to the center of the top of the tank. No other
type of tank is equipped with this type of walkway or control arm.
The tanks on the left end of row D are similar but smaller. The remain-
ing three tanks of row D are of the fixed-top (cone roofed) variety used
22
-------
to store less volatile products such as kerosene and jet fuel. An iden-
tifiable feature is the venting fixture located on the top of the tank.
The two tanks on the right are approximately 39 ft high and have a
diameter of 128 ft. The remaining tank is approximately 145 ft in
diameter and 43 ft high.
The two tanks at the top of the image, marked E, are the largest in the
immediate area. They are approximately 158 ft in diameter and 48 ft
high and are the only tanks in the photograph that employ above ground
piping in the revetted area.
The protective revetments are generally 5 to 6 ft high, and are in
good repair. From the stereoscopic analysis, the volume of the diked
area is sufficient to hold the volume of the tanks under consideration.
The enclosed area around the tanks appears to be well maintained and
includes service roads to each tank. These areas all show some signs
of vegetation and no distinguishable accumulation of spillage. No
trash or debris can be found in the enclosed diked areas. In general,
the tank farm would not constitute a potential threat even if it were
located adjacent to an inland waterway.
Multiband flights were flown over a bulk storage area for asphalt and
residual fuel oil. Asphalt and residual fuel oil storage can be
identified by black, sealed-roofed, round tanks. A ground truth
photograph of such a tank obtained on an initial ground truth mission is
shown in Fig. 3a. Figure 3b is a 70 mm aerial color photograph of the
asphalt bulk storage area. This area is also located approximately
one mile from the major waterway and does not directly represent a
potential or real threat. The value of the multiband flights in this
area was in the positive identification of asphalt spills through con-
trast enhancement. These results can be applied to other areas adjacent
to an inland waterway in order to determine whether an undefined spill,
and thus the spill source, constitutes a real or potential threat.
Positive ground asphalt spills are noted in areas A, B and C. Ground
truth teams had identified a leaky asphalt tank in this area. Figures
3c to 3f are photographs of the bulk storage area recorded with
film/filter combinations 2403/32, 2403/35, 2424/35, and 2424/99, respec-
tively. There appears to be little difference between the use of filter
32 and filter 35 with either film type 2403 of type 2424. A look at the
spectral responses of these filters shows that filter 32 passes more of
the visible spectrum (300 to 520 nm) than filter 35 (320 to 460 nm).
Since oil and asphalt have a strong reflectance in the ultraviolet and
blue spectral regions, the asphalt imagery obtained with filter 32 is
expected to be slightly darker than that obtained with filter 35 as
evidenced in Figs. 3c and 3d. A comparison of Figs. 3d and 3e reveals
that better asphalt-to-background contrast is achieved with film/filter
combination 2424/35 than with 2403/35. This is due to the near-infrared
(700 to 850 nm) background return recorded by film type 2424.
The maximum asphalt-to-background contrast is observed on imagery recorded
with film/filter combination 2424/99 as shown in Fig. 3f. Although
23
-------
a) Ground truth photograph
b) 2448/HF3
c) 2403/32
d) 2403/35
e) 2424/35
f) 2424/99
Figure 3 Asphalt bulk storage area
24
-------
definition or detail is lacking in the spilled areas, the asphalt appears
very dark against a light background. Since the 99 filter only passes
the yellow portion of the spectrum (500 to 600 nm), any return from
the asphalt in the ultraviolet and blue region of the spectrum has been
suppressed. Filter 99, however, does pass the 700 to 850 nm radiation
to which film type 2424 is sensitive. Because the asphalt appears very
dark in this imagery, it was concluded that asphalt reflects little
radiation in this spectral band. The contrast of asphalt to the back-
ground has been increased by suppressing the asphalt reflected radiation.
This phenomena has been termed negative contrast enhancement. It was
therefore concluded that the film/filter combination 2424/99 best detects
asphalt against the ground background. Film/filter combinations
2424/35, 2403/35 and 2403/32, however, give better definition of asphalt
spilled in water.
In the oil refinery bulk storage area, highly volatile gasoline products
are stored in white floating top tanks. The gasoline is treated with
caustic soda to improve the odor and to stabilize the final product.
The gasoline can acquire a variety of colors, depending on the additives
and amount of volatile material removed. On a particular ground truth
mission, gasoline that had a reddish appearance, was spilled in a diked
area as shown in Fig. 4a. A color aerial photograph of this area covered
during the multiband flights is shown in Fig. 4b. The dike in which this
gasoline spill is observed is marked A. This area is also located approx-
imately 1.2 miles from the river and does not represent a threat to the
waterway. The value of the multiband flight in this area is the determi-
nation of a film/filter combination that allows positive identification of
gasoline. These results are directly applicable to areas adjacent to an
inland waterway. An undefined spill can be positively identified as
gasoline and the source sought in the imagery to determine if a real or
potential threat to the waterway exists.
The two best film/filter combinations for gasoline detection were found
to be 2403/99 and 2403/65, as shown in Figs. 4c and 4d, respectively.
Film/filter combination 2403/65 imagery shows better contrast enhancement
of this area than 2403/99. Again, negative enhancement is emphasized by
suppressing the gasoline reflected radiation and making the area of inter-
est appear dark. Even though filter 65 passes a broader portion of the
spectrum (440 to 580 nm) than filter 99 (500 to 600 nm), it is more effec-
tive in the negative enhancement technique as it blocks out the near-red
spectral region.
Titanium Plant Bulk Storage
The titanium plant has many tanks which are primarily used for processing.
The larger tanks shown in Fig. 5 are for bulk storage. In addition to
these tanks, a large fuel oil tank is located to the south of the
facility. The tanks marked A in Fig. 5 are for liquid sulfur storage.
These three tanks are easily identified by the heating tube tops which
appear as light dots around the tank circumference. A walkway around
25
-------
a) Ground truth photograph
b) 2448/HF3
c) 2403/99
d) 2403/65
Figure 4 Gasoline bulk storage area
-------
the tank tops and an interconnecting walkway between tanks are also
evident. The tanks measure approximately 32 ft from the ground to the
base of the dome and have a diameter of 41 ft. The volume of these
tanks is estimated at 7500 bbls. The tanks appear well cared for and
no seepage or leakage were observed. A ground truth team learned that
a 3 ft high dike, not evident in Fig. 5, is located adjacent to the
river along the plant boundary line. There are no control dikes around
the tanks which would drain to the river if ruptured. Molten sulfur,
however, would solidify at ambient temperature and impede its own flow.
The effectiveness of the sulfur solidification as a protective measure
is unknown. This area, therefore, represents a potential spill threat
to the inland waterway.
Figure 5 Baseline imagery of titanium plant tank storage
27
-------
The two tanks labeled B are used to store sulfuric acid. They measure
approximately 54 ft in diameter and 32 ft from the ground to the base of
the domed top. Their volume is estimated at 13,000 bbls. each. These
tanks appear well cared for, but are unrevetted and would drain to the
river if ruptured. The ground truth team also learned the tanks are
connected through pipeline to four additional acid storage tanks within
the plant should they leak or rupture.
The tank labeled C is used to store sodium hydroxide and is approximately
35 ft in diameter and 24 ft high. The tank volume is approximately
4100 bbls. The tank appears well cared for, but is undiked and would
drain to the river if ruptured. A ground truth team learned that this
tank is also connected to additional storage tanks within the plant
should it leak or rupture.
The tanks labeled D, are used for fuel oil storage. The larger tank
measures 50 ft in diameter, 23 ft high, and has a capacity of 8000 bbls.
It is in good condition and is enclosed by a 5 ft high dike which is
adequate to hold the tank's contents. The area is clean and appears to
have foliage in part of the diked area. The smaller tank is approxi-
mately 35 ft in diameter, 24 ft high and has an estimated volume of
13,000 bbls. This tank appears well cared for and is enclosed by a 5 ft
high dike sufficient to hold the tank's contents.
Baseline flights over bulk storage areas were effective in determining
the number of tanks, the material in the tanks, the structural condition
of the tanks, the use of control dikes, the type of piping in these areas
the presence of debris in diked areas, and any seepage or leaks around
tank storage.
The value of multiband imagery is in establishing the spectral charac-
teristics and identifying the material that has been spilled or leaked
in the storage area. Once the material has been identified, the source
can be located and the potential or real threat to the waterway deter-
mined. In the titanium industrial area, no spills were found in the
bulk storage area on the multiband and color imagery. The effectiveness
of multiband imagery in this area could not be positively confirmed.
Color photography was concluded to be an essential part of the aerial
surveillance system for the detection of spills and spill sources.
Chemical Plant Storage
The part of a chemical plant shown in Fig. 6 is a storage area dominated
by the benzene storage tank. The floating top tank is approximately
105 ft in diameter and 30 ft high. In this photograph, the floating
tank top has descended approximately half way down the tank. The tank
volume is estimated at 52,000 bbls. The tank is surrounded by an 8 ft
high cement dike enclosing an area 286 ft long and 153 ft wide. The
area is clean and well tended although there was evidence of a liquid,
which could be water, on the ground. One major and one minor pipeline
are connected to the tank and run above ground.
28
-------
Figure 6 Baseline imagery of chemical plant tank storage
The river is located 4200 ft from the bottom of the photograph and the
terrain is flat with no specific drainage pattern. In addition, the
chemical industry is located behind a protective levee and therefore
does not represent a direct threat to the inland waterway. Drain gates,
which connect directly to the river through storm sewers, are evident
throughout the chemical plant and are used for the emergency disposal of
spilled hazardous materials. A close examination of such an area dis-
closes an indirect potential threat to the inland waterway.
Multiband imagery only has value in determining the type of spills and
in helping to better define drainage patterns. Over the chemical facil-
ity, however, very few spills were located (one is described under
loading facilities). The effectiveness of multiband photography could
29
-------
not be readily assessed in this area. Generally it appears that these
spills are easily defined with color photography and a knowledge of the
type of material stored in specific tank types.
Steel Mill Waste Storage Lagoons
In the steel industry, waste storage lagoons contain a variety of
materials. Such materials include lubricating oil from the rolling
mill cooling system, rust or iron which is acidicly removed from stored
steel materials, and water from various cooling towers. The cooling
water contains chemicals that are used to eliminate scale and fungus
from the cooling towers. The lagoon that contains the above materials
is treated with caustic soda to neutralize the acid and settle out solids
(iron scale) and is then passed through a filtration pond before it is
exited into the river.
The lagoon shown in Fig. 7 is located within the plant area of a large
steel mill. The liquid flow is from the right of the imagery to the
left. The lagoon was made by excavating four parallel trenches approxi-
mately 350 ft long, 70 ft wide, and 40 ft apart. The drain is connected
to the lower end on the first trench. Starting at the right, number one
and two trenches are connected at the top, number two and three trenches
at the bottom, and number three and four trenches at the top. The drain-
off trench is located at the lower end of trench number four.
At the time the image was recorded, number one and two trenches were
connected at their midpoint with the upper portions dammed off from the
lower portion. This was probably done to facilitate the removal of
sediments from the trenches. Footbridges and a chemical additive proces-
sor can be seen at the lower right end of the complex. At a later date,
other sections of the system will probably be closed off for maintenance
Multiband flights were conducted over the steel mill waste lagoons to
determine the best film/filter combination for material identification.
These results could then be applied to similar areas located adjacent to
inland waterways to identify and thus locate potential and real spill
sources threatening the inland waterway. The details of the multiband
analysis are given in Appendix C. For this area, color photography
revealed more information than any of the multiband images.
From the final steel mill drainage lagoon, the waste is channelled into
an impounding area in which other industrial complexes also dump waste
materials. Identification of all the dumped materials was not possible
because of the many companies involved. From the impounding area, the
material drains by a creek or canal to two sixty inch mains located
beneath the river dike, which empty directly into the Mississippi
River. This drainage constitutes an actual or real threat to inland
waterway. A valve in this main prevents water from backing up into the
impounding area during high water. A ground truth photograph of the
canal is shown in Fig. 8a. An aerial color photograph of this canal
30
-------
Figure 7 Baseline imagery of steel plant waste lagoon
before it reaches the river levee is shown in Fig. 8b. Figure 8c is an
aerial color image of the effluent passing under the levee, and Fig. 8d
shows the dissipation of the waste in the river.
The drainage pattern of this area was determined from the baseline image
shown in Fig. 9. The waste flow begins from the upper left of the
photograph and extends to the lower right. Additional drainage channels
can be observed in the upper right of the image. The entire area is
behind a flood control dike which runs parallel to the river. The area
is covered with revetted waste lagoons which are not presently being
used. Some foliage has returned to the area. A solid waste dump can
be seen in the dark areas in the upper right of the photograph.
-------
a) Ground truth photograph
b) Drainage to river
c) Effluent under levee
d) Waste dissipation in river
Figure 8 Steel mill and adjacent industrial waste drainage
In Fig. 8, the black color of the waste Indicates the presence of oil.
The multiband imagery confirms the nature of waste material. Multiband
photography is valuable for identifying the effluent material to deter-
mine if a real threat to the waterway exists. The best contrast was
obtained on imagery recorded on film/filter combinations 2403/99,
2403/35, 2424/99 and 2424/35. The images obtained are shown in Figs.
lOa, lOb, 10c, and lOd, respectively. These same film/filter combina-
tions were found to be effective for oil and asphalt detection. In
examining Figs. lOa to lOd, one concludes that filter 99 with both
film types 2403 and 2424 gives better contrast of the oil to background
than filter 35 because of its suppression of the oil reflectance in the
ultraviolet, blue, and red portions of the spectrum. Here again, negative
contrast enhancement is emphasized. Filter 32, which transmits these
32
-------
Figure 9 Baseline imagery of steel plant and adjacent industrial
waste drainage
spectral regions, results in reduced image waste background contrast
but increases the detail. Film/filter combination 2424/99 gives better
contast of the drainage area to the background than film/filter combi-
nation 2403/99 because it records more near-infrared radiation reflected
from the ground and vegetation. However, in the detection of oil in
water, the near-infrared absorption of water reduces the oil water
contrast.
13
-------
a) 2403/99
b) 2403/35
c) 2424/99
d) 2424/35
Figure 10 Steel mill and adjacent industrial waste drainage
-------
Figure 11 is a thermal infrared image of the drainage pattern observed
near the steel mill discussed above. The drainage creek is shown as a
white area, indicating that its temperature is higher than the background.
A temperature increase could be the result of the temperature at
which the waste is discharged from the industrial area, or of the absorption
and reradiation properties of the waste material. The value of thermal
infrared imagery is the detection of pollutant materials warmer than
their surroundings.
Figure 1 1 Thermal infrared image of steel mill and adjacent industrial waste drainage
Power Plant Fly-Ash Pond
Another type of waste lagoon identified during this project is the fly-
ash pond located adjacent to a power plant. Figures 12a and 12b are
ground truth and aerial photographs of this fly-ash pond, respectively,
The fly-ash residue from the boiler is transported to the pond by
water. The water flows into the pond at the dark area and settles out,
-------
By the time the water moves to the overflow on the opposite side of the
pond, which empties into the river, almost all solid matter has settled
out. The pond does not constitute a threat to the inland waterway.
The pond covers approximately 34 acres, with less than one-third being
covered by water level deposites of fly-ash.
Generally, ash piles are a black to dark gray color. The maximum
contrast of the ash pile to the pond was obtained on imagery recorded
with film/filter combination 2403/99 as shown in Fig. 12c. The ash
pile appears very dark as filter 99 suppresses the reflected radiation.
For comparison, the imagery obtained with film/filter combination
2424/65 is shown in Fig. 12d. As can be seen, the fly-ash pile pond
contrast is not as great as that obtained with film/filter 2403/99.
This is a result of water absorption in the near-infrared region
recorded by film type 2424. The advantage of the multiband imagery
is in identifying drainage patterns in the pond that are not apparent
in color imagery.
Waste Treatment Ponds
Water waste occurring in an oil refinery is pumped into one end of inter-
connecting lagoons where it undergoes biological oxidation. The waste
contains a small percentage of oil which accumulates in the ponds. The
heavier accumulated oil waste is removed from these ponds and trucked
to oil waste areas. Phenols and other oil waste products are broken
down by microorganisms which feed on these materials.
During the filtration from one pond to another, the oil water mixture
is aerated to increase the biological action. Algae are attracted to
the carbon dioxide generated by the bacteria, which completes the biolog-
ical process.
The oil detention pond area shown in Fig. 13 is part of a larger pond
complex. The complete complex is on the river side of a levee. A flood
control drain located north of these ponds reduces the potential threat
of river contamination by these ponds during times of high water. The
ponds themselves are enclosed by lesser dikes to ensure their separation.
The dikes appear well maintained and have controlled flow from one pond
to the next. The larger pond has an area of 6.8 acres but the depth is
unknown.
Because ground truth teams had positively identified oil in water,
extensive multiband flights were conducted over this area to determine
the best film/filter combination for the detection of oil in water.
These results are directly applicable to other areas for positively
identifying an unknown spill or seepage into an inland waterway as oil.
36
-------
a) Ground truth photograph
b) 2448/HF3
c) 2403/99
d) 2424/65
Figure 12 Power plant fly-ash pond
37
-------
Figure 13 Baseline imagery of waste oxidation ponds
A ground truth photograph of the pond is shown in Fig. 14a. An aerial
color photograph of this area covered during the multiband flights is
shown in Fig. 14b. The five "starred" areas are the aerators. The oil
waste initially pumped into the pond may be seen in the left-hand portion
of the photograph labeled area A. The best film/filter combinations for
identifying the oil in the ponds were found to be 2424/99, 2424/32, 2403/99,
and 2403/35, as shown in Figs. 14c, 14d, 15a, and 15b, respectively.
Similar results were obtained with filters 99 and 98 and with 32 and 35
with both film types 2403 and 2424. The maximum contrast of oil-water was
obtained with film type 2403 because of the strong absorption of water in
the near-infrared as recorded by film type 2424. Film type 2403 and
filter 35 or 32 and 98 or 99 give the best oil-water contrast.
38
-------
a) Ground truth photograph
b) 2448/HF3
c) 2424/99
d) 2424/32
Figure 14 Waste oxidation ponds
In this pond area, a similar contrast enhancement for oil-water imagery
is obtained with film/filter combinations 2403/99 and 2403/35. Since
filter 35 transmits the ultraviolet, blue, and red radiation reflected
from oil waste, the imagery obtained from 2403/99 was expected to be
darker than that obtained with 2403/35. The similarity of these images
indicates water absorption of this radiation.
An interesting result was observed in the area labeled C in Fig. 14b.
In the infrared imagery recorded on film type 2424, area C appears dark
while the pond to which it is connected is much lighter. On the film
type 2403 imagery, however, area C has the same tonal quality as the rest
of the pond. While this contrast enhancement is indicative of oil in
the infrared imagery, none appears visible in the color photograph. It
39
-------
was further noted that the area C image density is the same as that of
the pond containing the aerators. Ground truth revealed area C was
much deeper than the rest of the pond to which it is connected. Thus,
the darker pond area in the infrared imagery is a result of greater
near-infrared absorption by the deeper water. Film/filter combinations
2424/32 or 2424/99 may have some value in determining water pond depths
but the effectiveness and range of detectable depths is presently
unknown. Thus, one has to be careful in interpreting pond imagery on
film type 2424 so as not to mistake oil for water.
The imagery of this pond recorded with film/filter combination 2403/18A
is shown in Fig. 15c. Filter 18A transmits the ultraviolet radiation o£
the spectrum (300 to 400 nm). With this filter, some water penetration
can be seen since flow patterns from one pond to the next could be seen.
Since oil is strongly reflective in this spectral region, the observed
striations are believed to be oil.
a) 2403/99
b) 2403/35
c) 2403/18A
Figure 15 Waste oxidation ponds
40
-------
Oil Refinery Effluent
The waste treatment ponds described in the last section eventually empty
into the river and could represent a potential threat to the river. A
baseline photograph of the effluent area is shown in Fig. 16. The drain,
labeled A in the photograph, is approximately 28 ft wide and is controlled
by a large gate valve. From this point it flows underground to the river.
The river opening is approximately 15 ft wide and can be located by the
plume of liquid being dumped into the river.
Figure 16 Baseline imagery of oil refinery effluent
-------
An aerial color photograph containing the oil refinery effluent (labeled
A) is shown in Fig. 17a. A ground truth photograph of this effluent is
shown in Fig. 17b. Ground truth teams reported no visible evidence of
oil at this effluent. The white caps seen in both Figs. 17a and 17b are
an unusual feature of this effluent which the oil industry has been unable
to remove. Fish are known to gather at this effluent to feed.
Multiband photographs were taken of this area to determine if previously
successful film/filter combinations for oil in water detection would
detect any presence of oil at this effluent. An examination of the multi-
band imagery revealed the effluent image density was approximately the
same for all filters combined with both film types 2403 and 2424. This
was expected as the white foam reflects all visible radiation uniformly.
The white effluent recorded on film type 2424 had better contrast against
the water background than that observed with film type 2403. This is a
result of the near-infrared absorption of water recorded by film type
2424.
Filters ISA, 47B, and 39 transmit blue and ultraviolet radiation which
penetrates water surfaces. As pointed out in the last section, images
recorded with these filters and both film types 2403 and 2424 revealed
oil flow patterns in the lagoon. Figure 17c is the imagery of the oil
pond effluent as recorded by film/filter combination 2424/4715. The 47H
filter does allow some water penetration as the white foam can be seen
dissipating in the river.
While some contrast enhancement and water penetration are gained with
multiband photography, no oil could be detected in the water. Color
photography was concluded to be more effective for monitoring this type
of effluent.
Titanium Plant Effluent
Along the 1032 ft river front shown in Fig. 18, nine sources of water
effluents are detectable and represent a real or actual threat to the
inland water way. The pipes drain directly into the river with no
visible evidence of plant control facilities. The materials being
discharged appear the same on the baseline imagery. Ground truth of the
area disclosed the following information:
Source A - Undissolved ore-gangue from clarification tanks,
brownish black in color
Source B - Cooling basin overflow river water
Source C - Cooling and processing water from ore digestors
stack scrubbers not suited to recirculation
Source D - Process water which is orange in color
42
-------
Source E - Process and runoff water from plant, three separate
outlets
Source F - Main plant waste containing titanium dioxide, ferrous
sulfate, and sulfuric acid
Source G - Cooling water from sulfuric acid manufacturing process
mm
A ^^^
b) Ground truth photograph
a) 2448/HF3
c) 2424/47B
Figure 17 Oil refinery effluent
43
-------
The tanks identified in Fig. 18 are intricate components of the
manufacturing process and receive periodic inspection and the same
maintenance as any other manufacturing machinery. The tanks are partiali
masked by associate structures, but appear unrevetted and would drain to
the river if ruptured.
Figure 18 Baseline imagery of titanium plant effluents
44
-------
Multiband photography was flown over the titanium plant effluents to
determine if any film/filter combinations could be used effectively for
identifying materials discharged into the river. Figure 19 is an aerial
color photograph of the titanium plant and its effluents. The effluents
on the 70-mm photographs are distinguishable to the unaided eye, but a
10X magnifier helps to identify each area. Effluent A is brownish black
in color and consists of undissolved ore called ore-gangue. While no
active effluent is evident, the stagnant water in this area on the 70-mm
film displays a grayish color. During periods of low water the effluent
is piped further down the bank and exited at the water intake tower.
Effluent B primarily contains cooling water. On the color photograph
the water appears similar to river water. Effluent C is also cooling
and processing water. On the color photograph the effluent is lighter
than the river and is probably due to aeration. Effluent D is processing
water that contains iron sulfate from water treatment compounds from
boiler water treatment and appears orange in Fig. 19. Effluent F, the
main plant sewer, contains such wastes as titanium dioxides, ferrous
sulfate and sulfuric acid. The exited material appears very gray.
Effluent G is the raw river water used for cooling in the sulfuric acid
manufacturing area and appears light gray.
Figure 19 Titanium plant effluents
A.11 of the multiband imagery of these effluents looked approximately the
same. Therefore no one film/filter combination was effective in contrast
enhancing any of the effluents. This is the result of the spectral
characteristics of these effluents being either very broad or narrower
than the spectral bands achieved with the film/filter combination.
Overall, the color photography was very effective in determining the
effluents and their general spectral signatures.
45
-------
Figure 20 is a thermal infrared image of the titanium plant and its
effluents. Most of the plant processing areas appear very bright, indi-
cating a temperature increase over the background. Some of the effluents
also appear bright and represent a warm or hot effluent. From the ground
truth missions, process cooling water is known to be emptied into the
river and could constitute an infrared return. For this type of indus-
try, the value of infrared photography over multiband or color photog-
raphy is in identifying various processing facilities by their tempera-
ture signature and in identifying real thermal pollution threats to the
inland waterway.
Figure 20 Thermal infrared image of titanium plant and effluents
Sewage Effluent
In concurrence with the sites selected with the Environmental Protection
Agency, a sewage effluent outlet was included in the baseline and multi-
band flights. This was not a primary target but demonstrates the photo-
graphic detection of a potential threat (if the sewage to water ratio
is not in an acceptable range) and demonstrates the use of multiband
photography to distinguish the presence of a variety of materials through
their spectral characteristics.
46
-------
The sewage discharge, confined between flood dikes to form a simple
drainage pattern, is shown in Fig. 21. The drain is approximately 20 ft
wide and dumps into a navigation canal 468 ft wide. The effluent from
the sewage facility is very dark, indicating the presence of oil which
could come from the various industries, including a steel mill, which
drain directly into the sewage system. Figure 22 is an aerial color
photograph of the sewage effluent. The best film/filter combination for
detecting the effluent was found to be 2403/99, as shown in Fig. 22b.
The use of this film/filter combination confirms the presence of a high
percentage of oil. The film/filter combination suppresses the radiation
reflected from oil and again emphasizes negative contrast enhancement.
Figure 22c shows the imagery obtained with the film/filter combination
2424/99. The high absorption by water of near-infrared radiation reduces
the sewage-water contrast drastically. When the imagery is compared
with that obtained from the oil refinery, the observed contrast reduc-
tion is probably a result of near-infrared reflectance from materials
Figure 21 Baseline imagery of sewage plant effluent
A 7
-------
other than oil. The multiband imagery obtained with film/filter combi-
nation 2424/65 is shown in Fig. 22d. A contrast reversal of the effluent
is observed in the imagery. The near-infrared absorption of the water
provides a dark background for the lighter effluent. The use of filter
65 indicates a reflectance of the effluent in the 450 to 550 nm spectral
region.
a) 2448/HF3
b) 2403/99
c) 2424/99
d) 2424/65
Figure 22 Sewage plant effluent
48
-------
From a comparison of the imagery obtained with these three film/filter
combinations, it can be concluded that other waste materials besides oil
are present. The large number of industries dumping waste into the
sewage system precluded identifying specific waste products.
Power Plant Effluent
The power plant pumps water from the river into the main plant to be
used for cooling. The heated water plus the plant sewage constitute
the power plant effluent. Hence the threat to the inland waterway is
thermal pollution. The liquid discharge area of a power plant is shown
in Fig. 23. The primary discharge is through two cement structures
approximately 13 ft by 26 ft and equipped with gates. A secondary dis-
charge is located to the lower right of the cement structures and comes
from a cement pipe approximately 2 ft in diameter. The discharge con-
tains raw sewage and drainage from within the plant itself.
Figure 23 Baseline imagery of power plant effluent
49
-------
A ground truth photograph of the power plant effluent, Fig. 24, shows
that the spectral signature is not any different than that of the sur-
rounding river water. Hence multiband photography is of no value in
determining the type of material exited at this particular point. The
presence of the effluent and its location relative to the river can be
obtained from baseline photographs.
Figure 24 Ground truth photograph of power plant effluent
As the water exited at the power plant has a higher temperature than
the surrounding river, thermal infrared imagery should be effective in
defining this effluent. Figure 25 is a termal infrared image of the
effluent obtained from the Reconnaissance Laboratory Data base in which
the heated effluent does appear lighter than the river background.
Therefore, thermal infrared imagery is very effective in locating and
tracing a heated effluent that cannot be detected by multiband or color
photography.
Lime Sludge and Oil Waste Area
One non-oil waste material that may be found in abundance in an oil
refinery is lime sludge which is derived from the water softening pro-
cesses employed. Generally,! the lime sludge is trucked or pumped to a
land waste area for natural absorption and decomposition.
Much of the oil waste encountered from tank storage in the oil industry
is deposited in waste ponds. These differ from lagoons in that they are
stagnant and are not filtered and discharged into the river. Natural
biological action is the process used to break down the oil waste.
The large dump area shown in Fig. 26 is used for the disposal of lime
sludge and oil waste. The area is located 2.8 miles from the river and
does not represent a direct threat to the inland waterway. The non-
river location was covered in the baseline and multiband photography to
demonstrate the ability to establish drainage patterns and to positively
identify lime sludge. These results can then be applied to evaluate rea
and potential threats of similar sites located adjacent to inland water-
ways.
50
-------
Figure 25 Thermal infrared image of power plant effluent
The pit shown in Fig. 26 measures approximately 1018 ft at its longest
dimension and is 544 ft wide. The waste, by area, is approximately
one-third liquid. The area in the photograph slopes from right to left
and forms one portion of the low hills which surround this general area.
The waste is dumped into the upper right-hand corner of the pit and runs
to the lower left, as is evident by the drainage pattern. A dike, 18 ft
above ground level, is located to the left of the lime sludge dump area.
Although no seepage is observed on the imagery, a drain pipe can be seen
extending through the dike in the lower left-hand portion of the waste
area. Drainage terminates into a swampy area located to the left of the
lime sludge area.
The value of the multiband photography is the positive identification
of lime sludge. Such results can be applied to identifying an unknown
waste material to determine whether or not a waste area constitutes a
real or potential threat to an inland waterway.
-------
Figure 26 Baseline imagery of oil refinery waste storage
Figure 27a is a ground truth photograph of the lime sludge area while
Fig. 27b is an aerial color photograph of the same area covered during
the multiband flights. The lime sludge area is labeled A in the latter
photograph. The best film/filter combinations for identifying the area
were 2403/65 and 2403/99. These images are shown in Figs. 27c and
27d, respectively. Filter 99 transmits the yellow portion of the spec-
trum (510 to 600 nm) while filter 65 transmission is centered around the
blue-green spectral region (440-580 nm). Filters 98 and 75 yielded
almost equivalent images with film type 2403 as the images included in
this report. Filter 98 transmission is centered around 440 nm (390 to
500 nm) while filter 75 transmission is centered around 490 nm (450
to 540 nm). A good reflectance over the whole portion of the visible
spectrum was expected because of the lime sludge's whitish color.
-------
a) Ground truth photograph
b) 2448/HF3
c) 2403/65
d) 2403/99
Figure 27 Lime sludge waste area
The imagery obtained with filter 65 appears to have slightly better con-
trast than that obtained with filter 99 due to the greater spectral band-
width of filter 65. It should be pointed out that a positive contrast
enhancement technique was used here. The image contrast has been
optimized by choosing filters which transmit rather than suppress
electromagnetic radiation reflected from the target of interest. Film
type 2424 seemed of little use in this area as the near-infrared
reflectance of the background reduced the contrast of the lime sludge
area. In general, the multiband imagery defines the drainage patterns
better than color photography. However, this gain is so slight that
color photography appears to be the most effective detector of lime
sludge areas and the drainage patterns that occur within these areas.
53
-------
Oil Waste Dump Area
The area below the pipeline, located at A in Fig. 28, is a portion of
the waste disposal facility for a large oil refinery. The waste area
shown is approximately 1550 ft long and 1008 ft wide at the widest
point and is located between the flood levee and the river. The area
is essentially flat with a major revetment surrounding the whole area
to prevent any drainage into the river. In addition, each liquid pond
is revetted and solid waste is dumped in a random manner further breaking
up any permanent drainage pattern. Even though this area is located
south of a control dam, a large flood would make the whole area a
potential threat to the inland waterway.
Multiband photography was taken to determine the best film/filter
combination for identifying and detecting the various dumped waste
materials. These results can be applied to the detection and identifi-
cation of unknown accumulated waste materials to determine if such
material is a real or potential threat to the inland waterways. An
aerial color photograph of this oil waste area is shown in Fig. 29a.
Each area has been labeled in Fig. 29a, and they will be sequentially
discussed. The best multiband imagery was obtained with film/filter
combinations 2403/99, 2403/35, 2424/99, and 2424/35, and is shown in
Figs. 29b, 29c, 29d, and 29e, respectively.
The area labeled A in the aerial color photograph is primarily a heavy
oil waste area that has a very black appearance. The imagery obtained
with filter 99 for both film types 2403 and 2424 is darker than that
obtained with filter. 35. As before, the contrast enhancement is a
negative process based on the use of filter 99 to suppress the ultra-
violet, blue, and red electromagnetic radiation reflected from the oil.
Consequently, imagery obtained with filter 35 has less contrast as it
transmits the portion of the spectrum reflected by oil. Also, the back-
ground of the imagery recorded on film type 2424 is lighter than that
recorded on film type 2403. This is a result of near-infrared back-
ground reflected radiation and the extended spectral range of film type
2424. The oil waste area recorded on film type 2424 shows an increased
relative contrast over that recorded on film type 2403.
The oil waste area labeled B in the aerial color photograph will now be
discussed. On both film types 2424 and 2403, the imagery obtained with
filters 35 and 99 have approximately the same contrast. This indicates
that the ultraviolet, blue, and red radiation reflected from this area
is negligible. Note, however, that the imagery of area B recorded on
film type 2424 is lighter than that recorded on film type 2403. It can
be concluded that area B reflects strongly in the near-infrared. Note
that these spectral characteristics are different than those of area A.
From these multiband photographs, the various waste materials can be
separated and identified. From a ground truth mission it was learned that
54
-------
Figure 28 Baseline imagery of oil refinery waste area
55
-------
area C in Fig. 29a is an eight-year-old waste area containing oil-coated
lime sludge. The brownish red color here indicates the pressure of
filter cakes. A filter cake is diatomaceous earth used for filtering
of certain motor oil additives in the processing plant. Upon saturation,
it is flushed with oil to remove the motor oil additives. It acquires
a blackish appearance and is dumped in an oil waste area. When the
filter cake is rained on, the oil is leeched out in a reddish brown
residue seen in area C. A ground truth photograph of a filter cake and
oil residue is shown in Fig. 29f.
The multiband image of area C recorded with film/filter combination 2403/99
is very dark compared to that recorded by 2403/35. Some detail of drainage
is apparent in the latter image. The results also support the reflectance
of the ultraviolet, blue, and red radiation from the waste area. An
examination of Figs. 29e and 29f of the imagery recorded with film/filter
combinations 2424/99 and 2424/35 further supports this conclusion. The
image of this area recorded with film type 2424 is lighter than that re-
corded with film type 2403, indicating a strong near-infrared reflectance.
Area C is similar to area B except that there is some ultraviolet, blue,
and red reflectance from this area.
Area D on Fig. 29a is the green and orange ponds that are very apparent
in the aerial color photograph. The green pond is a result of algae
that have formed on the oil waste product. The orange pond is an
emulsified oil-water area. The green pond imagery recorded with
film/filter combination 2403/99 is lighter than that recorded on 2403-35.
This can be anticipated as filter 99 transmits the yellow-green radiation
while filter 35 absorbs this radiation. The algae pond imagery has
approximately the same tone when recorded with both film/filter combina-
tions 2424/35 and 2424/99. The images are slightly lighter than that
recorded on 2403/35, indicating some near-infrared reflectance.
The orange region recorded with film/filter combinations 2403/99 and
2403/35 appears to have the same image density. This is expected as both
filters absorb in the orange spectral region. The imagery of the orange
area recorded with film/filter combinations 2424/99 and 2424/35 appears
to have the same contrast relative to the background. The area appears
slightly darker on film type 2424 than on film type 2403, indicating some
near-infrared absorption. In general, the multiband images of the orange
and green ponds tends to be somewhat confusing and more is gained from the
image recorded on color film type 2448.
Another oil waste area not located on the inland waterway is discussed
in Appendix D. The area is interesting from the multiband photographic
point of view because the oil waste is dispersed among stagnant water ponds
vegetation, filter cakes and emulsified oil waste. The conclusion for
this area is that film type 2403 is better for identifying oil waste
than film type 2424.
56
-------
a) 2448/HF3
b) 2403/99
c) 2403/35
d) 2424/99
f) Ground truth photograph
e) 2424/35
Figure 29 Oil waste area
GP71-1642-35
57
-------
Titanium Plant Open Storage of Raw Materials
In many industrial areas it is necessary to store large volumes of raw
materials on the ground. A portion of the storage facilities for a ti-
tanium plant are shown in Fig. 30. From the black-and-white photograph
the material identification is unreliable. A sulfur pile, labeled A in
Fig. 30, was identified on an initial ground truth mission. It has a
base of approximately 60 x 290 ft and is partially enclosed by a wall
4.5 ft high with access at either end for loading. The sulfur is stacked
10 ft above the wall. Since the entire area is protected by a levee,
located at the top of the photograph, the river is not threatened by any
runoff. If the levee were not present, however, natural drainage from
such a source located within a 100 yds of the river would constitute a
threat to the river. Additional items in the storage area labeled B,
C, and D are a reserve coal pile, scrap metal, and fuel oil tank,
respectively. The latter is enclosed by a protective dike, and a small
fuel oil spill is observed at the base of the tank. The top of the
tank appears to be stained from a previous overflow.
Multiband photography was taken of this area to determine the best
film/filter combination for sulfur detection and identification. The
results could then be applied to areas adjacent to an inland waterway
to detect and positively identify an unknown spill to determine if a
real or potential threat to the waterway exists.
A color aerial photograph of the area including the sulfur pile labeled
A is shown in Fig. 31. An examination of the multiband imagery revealed
that the similar contrast enhancement of the sulfur pile was achieved
with filters 98, 99, 65, and 75 with both film types 2403 and 2424.
Better contrast of the sulfur pile to the background was achieved with
film type 2403 since the detection of background reflected near-infrared
radiation by film type 2424 reduced the overall image contrast. Slightly
better contrast enhancement was achieved with filter 99 as its maximum
transmittance was in the yellow portion of the spectrum. Even with
multiband image enhancement, the detection of sulfur was optimized with
color photography.
Coal Storage
The commercial coal storage yard shown in Fig. 32 is located at the in-
tersection of two rivers. The coal pile is approximately 780 ft long
and 535 ft wide and can be serviced by barge, rail, or truck. The
storage is open without any protective dikes and apparently has no under—
ground drainage system. A natural runoff pattern exists to the
Mississippi River and to the river located at the right side of the
photograph. Coal dust in suspension can be detected in the river down
stream from the barge handling facility. Coal sediments can be seen in
the creek just below the railroad bridge in the lower right of the photo-
graph.
58
-------
"••--
Figure 30 Baseline imagery of open sulfur storage
Figure 31 Sulfur storage area
59
-------
Figure 32 Baseline imagery of open storage of coal
A color aerial photograph of the area is shown in Fig. 33. A close
examination of the area with a 10X magnifier revealed coal dust drain-
age from the coal pile into the river. Likewise, coal spilled in the
river during barge loading can also be seen. The multiband imagery was
examined to see if contrast enhancement of coal and coal dust drainage
was achieved with any film/filter combination. Some contrast enhance-
ment was obtained with filters 99, 98, and 75 with both film types 2403
and 2424. The contrast enhancement was obtained by suppression of the
reflected radiation from coal which occurs in the ultraviolet and blue
portions of the spectrum. After a detailed analysis was made of the
multiband imagery, it was concluded that more information was available
in the color photography. Making the coal pile darker in the multiband
imagery does not particularly aid the photointerpreter.
60
-------
Figure 33 Coal storage and loading area
Cement Plant Shale Quarry
In accordance with the sites selected with the Environmental Protection
Agency, a cement manufacturing plant was covered in the baseline and
multiband flights. The most interesting feature of this facility was
the plant shale and limestone quarry. Fig. 34 is a baseline photograph
of this quarry. It measures approximately 1000 ft at its longest point
and 600 ft at the widest point. The pit is being worked at one end
where the mechanical shovel and bulldozer are located. The material is
transported to the opposite side of the pit by truck where it is dumped
into a hopper. From there it is moved to the plant by an underground
conveyer belt. A power drill on the upper level of the active area
above the large shovel is used to bore shot holes for dynamiting.
Upon stereoscopic analysis, the quarry, located adjacent to the river,
was found to have no natural drainage. Hence the area does not repre-
sent a direct threat to the inland waterway. All drainage collects in
a pond located at the center of the quarry. The water is pumped from
there to the main plant, where it is combined with the river water used
to cool the kiln, for disposal into the adjacent river. Ground truth
teams found no detectable waste material such as shale or limestone in
the cement plant effluent.
The value of the multiband flights was in determining the best film/filter
combination for detecting shale and limestone. The results can be used
to locate and identify shale and limestone in other areas adjacent to
inland waterways to determine if a potential threat to aquatic life
exists.
61
-------
Figure 34 Baseline imagery of open shale quarry
A ground truth photograph and an aerial color photograph of the quarry
are shown in Figs. 35a and 35b, respectively. From the photographs,
the shale and limestone is observed to be grayish white. The multiband
imagery analysis revealed the best film/filter combinations were 2403/99
2403/98, 2403/75, 2424/99, 2424/98, and 2424/75. The imagery recorded
on both film types 2403 and 2424 with filters 99, 98, and 75 appeared vety
similar. Figures 35c and 35d are the imagery of the shale and limestone
62
-------
a) Ground truth photograph
b) 2448/HF3
c) 2403/99
d) 2424/99
Figure 35 Shale quarry
63
-------
quarry obtained with film/filter combinations 2403/99 and 2424/99
respectively. All of the 2403 imagery gave better contrast of the
shale quarry to the background reflected near-infrared radiation
recorded by film type 2424. The wide range of acceptable filters is
understandable as the quarry reflects visible radiation uniformly.
For this particular area, color photography yields the maximum amount
of information.
Pipelines Over or Near Waterways
Figures 36a and 36b are the baseline and ground truth photographs, respec-
tively, of a portion of the pipeline facility of a barge transfer system
used by a major oil refinery. This portion is connected to the main
refinery by four above ground pipelines. All lines appear in good repair
with no spillage visible. A small amount of spillage is evident around
a tank truck transfer point. The river transfer points are two floating
docks connected to the shore and equipped with overhead pipelines. These
floating docks could be a potential spill threat if a line were unattended
during the actual transfer operations. One small open pit is located at
the facility and shows signs of petroleum dumping. The drainage, however,
does not reach the river. Generally, the overall appearance of this trans-
fer area indicates good maintenance.
The petroleum transfer point shown in the base and ground truth photo-
graphs of Figs. 37a and 37b, respectively, is very similar to the transfer
area discussed above. The facility is connected to the main plant by
five pipelines and has pumping and flow regulation equipment on site.
Two revetted tanks on the shoreline were being improved at the time
the baseline photography was taken. The actual loading is done from
two floating docks moored to the shoreline and equipped with overhead pipe-
lines. The danger of spill threat during the actual transfer operations
also exists here. No spills are in evidence at the time of photography.
The value of multiband photography over areas containing pipelines over
or adjacent to waterways is in detecting and identifying spills occurring
from pipe leaks or during barge-loading operations. The two barge-loadings
shown in Figs. 36 and 37 were covered during the multiband flights. During
these flights, no pipe leaks or spills in the barge-loading operations
were detected. Hence, experimental evidence was not available to assess
the effectiveness of multiband imagery in this area. Samples of the
multiband imagery of one of these barge-loading areas are shown in
Fig. 29. Oil would be the most probable spill threat to occur in this
area. From other sections of this report the best film/filter combina-
tions for detecting this spill would be 2403/99, 2403/32, 2424/99, and
2424/32. Film type 2403 would be more effective over the harbor while
film type 2424 would be better over land areas. This result is based
on the near-infrared absorption of water that reduces the oil-water
contrast on film type 2424 imagery and on the ground reflection of
near-infrared radiation that enhances the oil-ground contrast in film
type 2424 imagery.
64
-------
Figure 36 a) Baseline and b) ground truth imagery of barge loading area
65
-------
Figure 37 a) Baseline and b) ground truth imagery of barge loading area
66
-------
Chemical Plant Loading Areas
In all the industrial areas covered during the flight plan, many examples
of truck and rail loading were recorded. Loading areas having a natural
drainage pattern to the waterway could represent a potential spill threat.
Figure 38 is a portion of a chemical loading facility covered during
the flights. While the area is behind the flood levee, the analysis
made here demonstrates the technique that can be used in similar areas
located on an inland waterway. Areas A, B, C, and D are a sulfuric acid
loading area, a sulfuric acid contact plant, cooling towers, and a
reserve coal pile, respectively. The chemical storage area has one
railroad siding equipped to load nine railroad tank cars sequentially
and one truck facility capable of handling two tank trucks simultaneously.
There are six large storage tanks, four smaller tanks, and two more
large storage tanks under construction. The large and small tanks have
a diameter of approximately 35 ft and 15 ft, respectively, and are
approximately 20 ft high. From the black-and-white imagery, there are no
apparent spill threats in the area. The area is very flat and would not
represent a potential threat even if it were directly adjacent to the
river. A very indistinct mottled appearance on the ground is attributed
to natural soil difference.
Generally, the number of spills detected in all the industrial loading
areas was very small. One example of a possible spill in the chemical
plant is labeled A in Fig. 39a. A ground truth team learned the
yellow residue on the ground at the sulfuric acid loading area is a
result of storage of bulk sulfur in past years. Multiband photography
was evaluated to determine if any film/filter combination was effective
in identifying the sulfur deposits. The results could be used in
areas adjacent to waterways to identify an unknown deposit to deter-
mine if it represented a real or potential threat to the waterway. An
examination of the multiband imagery revealed that film/filter combina-
tions 2403/99 and 2424/99 gave the maximum contrast enhancement. These
images are shown in Figs. 39b and 39c, respectively.
Imagery obtained with filters 98 and 75 for both film types 2403 and 2424
gave the next best contrast enhancement of the deposit. It is reasonable
to expect the maximum positive contrast enhancement with filter 99 as
its peak transmission in the yellow portion of the spectrum. The amount
of contrast enhancement is so slight that color photography is as
effective in identifying and detecting this deposit as the multiband
photography.
Oil Refinery Loading Area
The photograph shown in Fig. 40 is a portion of an oil refinery
having considerable loading facilities, tank storage, and administration
buildings. The area is located 1.6 miles from the river behind a
flood levee on a flood plain which has little or no natural drainage.
67
-------
Figure 38 Baseline imagery of chemical plant acid loading area
68
-------
a) 2448/HF3
b) 2403/99
c) 2424/99
Figure 39 Sulfuric acid loading area
69
-------
Figure 40 Baseline imagery of oil refinery loading area
70
-------
The area in Fig. 40 does not have a ten foot difference in elevation
between any two natural ground points. Any industrial complex located
on such a flat terrian will require man-made controllable drainage for
the removal of waste. Consequently the area does not represent a direct
threat to the inland waterway. The analysis used here, however, can be
applied to similar areas located adjacent to inland waterways.
The area marked A is for loading of railroad tank cars and contains
five dead-end sidings. A maximum of 70 tank cars could be serviced
from 35 double stations while 20 additional tanks cars could be serviced
from the remaining single stations. The petroleum spillage is quite
evident as an accumulation of very small spills rather than one large
spill. Because the land is flat, there is no evidence of spills
flowing out of the area requiring protective measures. Such a site
located as an inland waterway would constitute a potential spill threat
because of the large volume of petroleum products handled here.
The two areas marked B are for the loading of tank trucks. The area
adjacent to A would probably handle four trucks simultaneously and
shows some spillage accumulation which again is confined to this area.
The area B closest to the main road will probably handle six trucks
simultaneously and shows little evidence of spillage. This is probably
due to the cement drive which drains to an underground sewer system.
The areas marked C are administrative facilities while the building
marked D is a warehouse.
Multiband imagery of an oil refinery loading area allows identifying
spilled materials, and hence the spill source, by their spectral signa-
ture. Transferred materials in an oil refinery include final products
such as gasoline, oil, and asphalt, and also waste materials such as
lime sludge and crude oil waste. The best film/filter combinations to
detect these materials have been discussed previously. Film/filter
combinations 2403/99, 2403/32, 2424/99, 2424/32 have been effective in
detecting oil and oil waste products through contrast enhancement on the
multiband imagery. Similarly, film/filter combination 2403/65 is effec-
tive in contrast enhancing gasoline spills and lime sludge waste material,
These film/filter combinations should be effective in detecting real
and potential spill threats in an oil refinery loading area located
adjacent to an inland waterway.
Refineries and Industrial Processing Facilities
The industrial complex shown in Fig. 41 is the central area of a
petroleum refinery where considerable processing and manufacturing is
carried out. The area is of interest as the interconnecting pipeline
and processing facilities can represent potential spill threats.
Although this area is two miles from the river, the analysis used here
can be applied to areas located adjacent to an inland waterway.
71
-------
Figure 41 Baseline imagery of oil refinery
Although its complexity makes it a difficult area to analyze from aerial
photographs, the following labeled areas are easily defined.
A - Product Warehouse - Storage with railroad car and truck loading
facility
B - Overhead enclosed conveyer system
C - Railroad tank car loading facility
D - Tank truck loading facility
E - Liquid asphalt storage in dark tanks for heat absorbing quality
F - Cooling units with fans visible in top of units
72
-------
G - Catalytic Reformer for the production of high octane aviation
gasoline
H - Twin Catalytic Cracking Units for the refining of crude oil
I - Overhead pipeline complexes for the transfer of liquids within
the plants
The tracing of individual pipelines through the complex is impossible.
Although the amount of detail within the photograph is exceptional, the
complexity of the operation is beyond determination from aerial photo-
graphs and no potential spill threats could be determined.
The value of multiband imagery from a refinery and processing area is for
detecting and identifying pipeline and storage tank spills of oil and
hazardous materials which could be potential threats to an inland
waterway. Since no leaks from ruptured pipelines were observed during
the multiband flights, no multiband imagery could be evaluated for this
area. For an oil refinery, it can be deduced from previous sections
that film/filter combinations 2403/99, 2403/32, 2424/99, 2424/32, and
2403/65 would be effective in enhancing spills. Over other industrial
areas, however, color imagery was found to be more effective for detect-
ing spills than any of the multiband imagery.
Figure 42 is a thermal infrared image of an oil refinery processing
area that was obtained from the Reconnaissance Laboratory Data Base.
The value of thermal infrared imagery of this type of area is in
identifying storage tanks, pipelines, and processing facilities that
contain recently processed materials that are warmer than their
surroundings. From the imagery, pipelines containing temperature
elevated materials can be traced to their respective storage facilities.
Therefore, thermal infrared imagery can furnish information that cannot
be obtained from color or multiband imagery. Such additional information
can be of value in identifying particular industrial processes but suffi-
cient information for locating and identifying real and potential spill
threats can be obtained from baseline and multiband photography.
Evaluation of Oblique Photography
An oblique photograph of an oil refinery complex shown in Fig. 43 is
typical of this type of imagery. North is to the left of the photograph.
Detail is greatly degraded by ground haze in the upper 1/3 of the photo-
graph. To the layman, it is superior to vertical imagery because the
viewer has definite clues regarding relative object heights. To the
photointerpreter, however, the scale changes throughout the photograph,
making mensuration difficult. The same area of the photograph, where
detail is discernible, could be covered by a vertical photograph from
a higher altitude. The scale of such a photograph would be constant
throughout and contain more reliable vertical information when viewed
stereoscopically.
73
-------
Figure 42 Thermal infrared image of oil refinery
Effect of Weather Conditions on Imagery
During the flight test program, aerial imagery was taken only on clear
or slightly hazy days to establish the maximum effectiveness of an
aerial surveillance system. Haze and air pollution over the industrial
areas, however, did affect the resulting imagery. Figure 44a is an
aerial photograph of a chemical processing area obtained during condi-
tions of haze and heavy air pollution. Figure 44b is an aerial photograph
of the same area obtained on a clear day. Both of these images were
recorded with film/filter combination 2403/47B. Haze and air pollution
drastically affect the contrast and detail in the aerial photography.
Because of the limited number of clear days, the effect of adverse
weather conditions other than haze on multiband imagery needs to be
established. Knowledge of these effects on multiband imagery and
possible compensating adjustments on camera settings could improve the
effectiveness of an aerial surveillance system.
74
-------
Figure 43 Oblique imagery of oil refinery
Effectiveness of Temporal Change Detection
Detailed image analysis of the multiband imagery revealed the effective-
ness of an aerial surveillance system as a temporal change detector. For
the multiband analysis, the photointerpreter had to be careful to choose
imagery containing the same or similar information even though the flights
were made on different days. Imagery not containing the same information
points out the effectiveness of an aerial surveillance systems as a temporal
change detector. In Figs. 45a and 45b are aerial color images of steel-
waste lagoons photographed on sequential days. As can be seen, the
material in these lagoons has changed drastically from one day to the
next, indicating a high flow rate. From the multiband imagery it was
concluded that an aerial surveillance system can be an effective temporal
detector for determining flow rates, spill clean-up time, stagnant lagoons,
newly formed waste areas, and rate of usage of raw materials.
75
-------
Figure 44 Imagery recorded on a (a) hazy and (b) clear day
b)
Figure 45 Imagery of same area photographed on sequential days
76
-------
SECTION VII
GENERAL SUMMARY OF IMAGE ANALYSIS
The value of multiband photography is in detecting and identifying
spill threats and effluent materials through their spectral signatures.
It is imperative to identify oil and hazardous material spills before a
potential or real threat can be determined. Multiband photography is
also helpful in determining flow and drainage patterns in waste areas
and lagoons.
The best film/filter combinations for oil detection were found to be
2403/99, 2403/32, 2424/99, and 2424/32. The same filter was used with
both film types 2403 and 2424 because each film offers advantages for
particular backgrounds. Film type 2424 records the water near-infrared
absorption and reduces the oil water contrast. Thus, film type 2403 is
best for oil detection in large water areas such as oil-oxidation ponds.
It is also more effective than film 2424 in oil detection in waste areas
where water ponds and moist ground are intermingled with the oil waste.
The detection of near-infrared reflected radiation from water and ground
moisture by film 2424 results in imagery containing many shades of grey
that can be mistaken for oil. Film type 2403 reduces the contrast of
these water areas, thereby accentuating oil detection.
For a uniform ground background, however, the near-infrared reflected
background radiation recorded on film type 2424 enhances oil detection.
It should be emphasized that filter 99 is used to achieve a negative
contrast enhancement by suppressing the oil-reflected radiation. Sub-
stituting filter 32 for 99 results in sensing the ultraviolet, blue, and
red portion of the spectrum reflected by oil. While the corresponding
image contrast is not as great as that obtained with filter 99, more
detail of the oil distribution in water or on the ground is observed.
Similar results were obtained by substituting filter 35 for filter 32.
Filter 32, however, transmits more of the blue and red portion of the
spectrum than filter 35. Likewise, filters 98 and 75 yielded imagery
that was only slightly inferior to that obtained with filter 99. This
is expected as filters 98 and 75 begin to transmit in the blue spectral
region.
Other oil refinery materials investigated with multiband photography
were spilled gasoline, lime sludge waste, and an oxidation pond effluent,
The gasoline spill and lime sludge waste imagery were both enhanced
through use of film/filter combinations 2403/65 and 2403/99. Color
photography, however, was equally effective in detecting these materials,
The imagery obtained with filters 18A, 47B, and 39 showed some water
penetration as flow patterns were evident in the oil oxidation ponds.
Although this filter does record the ultraviolet radiation reflected
from oil, less oil water contrast is observed with this imagery than
with the imagery recorded with the film/filter combinations described
above. Similar imaging results for filters ISA, 47B, and 39 were ob-
served at the oil refinery effluent. Color photography is as efficient
in detecting the effluent as any of the multiband imagery.
77
-------
In other industrial areas, only slight contrast enhancement of spills
and effluents and raw materials was observed on the multiband imagery.
Thus, color photography was concluded to be the best detector for these
areas.
The baseline photographs of industrial areas were found to be effective
in identifying potential and real threats to inland waterways from
storage equipment, pipeline systems, tank farms, dike conditions, and
the presence of trash and debris in diked areas. In many cases, the
materials stored in various tanks could be determined from the tank
geometry and construction. Protective measures such as tank diking were
also determined from baseline photographs. While spill threats and
effluents are evident on the black and white baseline imagery, color and
multiband imagery are more effective in identifying the spill threats and
effluent materials. Baseline stereographic pairs were also used to
identify potential and real threats by estimating dike and tank heights
drainage patterns, arid the runoff pattern of openly stored raw materials
While the areas of lagoons and ponds could be approximated from base
photography, the lack of depth information prevented accurate determina-
tion of the lagoon and pond volumes. Accurate flow rates of ponds and
lagoons could not be determined from the baseline imagery.
During the image analysis, various techniques were employed to extract
the maximum amount of information from the imagery. These included
adequate choice of image scale for baseline and multiband flights,
stereographic image analysis, comparative coverage analysis, and ground
truth correlation.
The baseline flights were flown at altitudes ranging from 1300 to 3000
ft above the ground. For the 6 in. focal length Zeiss lens, the image
scale ranged from 1:2600 to 1:6000, respectively. From this imagery,
the necessary mensuration work was easily performed. The scale also
allowed identifying individual pipelines with the aid of 10X magnifier.
For the 9 in. base film format, ground coverage of 3000 ft was sufficient
to identify a large area of the largest industrial site of interest. This
simplified the construction of mosaics of the total industrial area. The
use of a smaller image scale would have increased the number of individual
frames and increased the complexity of mosaic construction.
Multiband imagery obtained from an altitude of 1500 ft above the ground
with the 50 mm Hasselblad lens had a scale factor of 1:9000. This scale
was found to be adequate for image detection of spill threats, effluents
waste lagoons, raw material storage, and waste areas. These areas were
easily identifiable in oil industry imagery while a 10X magnifier was
helpful in examining other industrial imagery.
Stereoscopic analysis of baseline images is necessary for determining
tank and dike heights, drainage patterns, and the volume of openly stored tt
materials. Stereoscopic analysis was found to help in material identi-
fication. Knowledge of the height of a waste area relative to the
78
-------
surrounding area can aid in determining the type of material in a par-
ticular area, A depresseion can be representative of a quarry while a
rise can be evidence of dumped waste.
Comparative coverage was found to be useful in positively determining
the best film/filter combination for material identification. This
technique compared the film/filter sections for material identification
in one area with those in another area. The use of this technique in
Section VI is obvious. Many more comparisons were employed than those
included in this report. It is also advantageous to compare dissimilar
areas. For example, in a bulk oil storage area, the film/filter effect-
iveness was determined by comparing oil and water spills.
Perhaps the single most valuable technique for determining the system
effectiveness was the correlation of ground truth information with base-
line and multiband imagery. As pointed out earlier, multiband flight
lines were selected where ground truth information and imagery were
available.
The use of the techniques discussed above has resulted in a surveillance
system that has maximum effectiveness for detecting real and potential
spill threats.
79
-------
SECTION VIII
EQUIPMENT, PERSONNEL AND PROCEDURES
FOR SYSTEM IMPLEMENTATION
The equipment, personnel, and procedures necessary to implement an aerial
surveillance system such as that employed during this project are dis-
cussed in this section.
The equipment necessary for an aerial surveillance system can be sub-
divided into three categories: flight equipment, film processing
equipment, and image analysis equipment. The flight equipment includes
the Zeiss mapping camera or equivalent, for baseline imagery and four
Hasselblad cameras or equivalent, for the multiband imagery. The latter
require additional accessories, such as filters, filter holders,
batteries, battery chargers,a command unit to trigger the four cameras
simultaneously, quick mounts, and at least 16 cassettes. Additional
magazines are not a necessity but do allow inflight camera reloading.
The specific Hasselblad lenses depend on the required scale and
flight altitude.
The film processing equipment for both 70-mm black-and-white and color
film includes 70-mm Nikor reel and tanks capable of handling 15 ft. of
film, a Pierce "ROK-IT" agitator to provide chemical agitation during
processing, and an Oscar Fisher dryer. In addition, the chemicals, such
as developer, fix and hypo-clearing bath, are needed for processing the
black-and-white film. Kodak E-3 chemical kits provide the color
processing chemistry. The use of these chemicals is described in
Appendix B. The 9-in. black-and-white and color film can be hand
processed but with considerable difficulty. A Versamat processing
system would allow the processing of both 9-in. and 70-mm color and black-
and-white photography. The Versamat, however, represents a large initial
investment, but greatly facilitates film processing. For a long-term
program, such a system would pay for itself. Without Versamat develop-
ment, additional general processing facilities are required, such as
temperature controlled water baths, thermometers, trays, and associated
glassware. It is also necessary to provide 70-mm and 9-in. reels and
cans. If the processing is to be performed commercially, only those
companies capable of performing precision processing should be considered.
A standard light table with 10X and SOX magnifiers and a stereoscope
are absolutely necessary for image analysis. An Air Force Height
Finder (parallax bar) is needed to determine drainage patterns and
estimate the heights of tanks, dikes, and waste piles. To do actual con-
touring of land areas, a paper print stereo-plotter is required.
Three or four personnel are required to implement an aerial surveillance
system independent of the pilot and any supervisory personnel. These
consist of one to two film processors, an aerial cameraman, and a
photointerpreter.
81
-------
The number of film processors will depend on the amount of film to be
processed and on the required turn around time. (During this project,
one person processed 70—mm black-and-white film, while another
processed the 70-mm color film. Versamat development was used only on
the 9-in. film. Two people were necessary because of the time factor
involved and the large quantity of exposed film.) These individuals need
a strong background in dark room and processing techniques.
An experienced aerial photographer is also required for obtaining the
aerial imagery. This individual needs to be well-versed on aerial camera
systems, and to be able to determine the camera exposure, the
eter setting to achieve the degree of overlap, load and unload the
and coordinate with the photointerpreter in determining the flight plan.
One of the functions of the photointerpreter is to determine the flight
plan with the aerial photographer. The flight plan includes the proper
aircraft headings, the number of exposures, and the number of flight
lines and altitude needed to adequately cover the target area. In
addition, the photointerpreter is also required to perform the image
analysis. At minimum, the photointerpreter should have the equivalent
knowledge of a military photointerpreter. This background enables him
to identify industrial storage, processing, and transfer areas, to per-
form the stereoscopic and mensuration analysis, and to correlate the
images with existing maps. At maximum, the photointerpreter should be
able to apply his analysis to the location and identification of
potential and real spill threats to inland waterways. This further
includes the identification of oil and hazardous materials on multiband
imagery and requires him to understand the use of spectral filters in
conjunction with film spectral sensitivity and the general laws of
reflection and absorption.
The procedures for attaining aerial imagery can be divided into two
tasks. The first task consists of obtaining a general survey of an
industrial area or areas for assessment of potential and real spill
threats to the inland waterways. Having chosen a specific industrial
site or sites, a baseline flight is flown with 9-in. black-and-white
film. Upon processing, the photointerpreter notes particular areas on
the baseline photography over which multiband imagery is required to
positively identify an unknown spill to determine whether it is a threat
to the inland waterway. Three film/filter combinations are chosen for
three of the Hasselblad cameras while the fourth camera contains color
film type 2448. The flight plan, including the area of interest, number
of exposures required to cover this area, flight lines, altitude, camera
settings and aircraft headings would be determined before the flight.
Upon completion of the flight, the film is processed and given to the
photointerpreter. He uses the multiband imagery to identify the mater-
ials and drainage patterns of the particular spill threats, waste areas
and effluents originally noted on the baseline photography.
The second task involves the collection of imagery of a specific pre-
determined area. In this case, both the baseline and multiband camera
82
-------
array are flown simultaneously. Under these conditions the baseline
camera would contain 9-in. color film while four film/filter combinations
would be contained in the Hasselblad array. The baseline photography is
still used for mensuration work but also provides the color imagery to be
used in conjunction with the multiband images. The procedure for
obtaining this imagery is essentially the same as above. The Hasselblad
magazines are limited to 70 exposures while the Zeiss camera can achieve
in excess of 110 exposures. Thus, the photographer would have to choose in
flight the multiband targets of interest. Hence the Hasselblad cameras
would have to be triggered manually while the Zeiss camera would be
automatically triggered by the intervelometer. Both film types would be
processed and given to the photolnterpreter for image analysis.
By excluding processing time and photographer manhours, the cost of a single
multiband flight can be approximated. First, assume that both the Zeiss
and four Hasselblad cameras are flown simultaneously. A 125 ft roll of
9-in. color film type 2448 costs approximately $100. A 150 ft roll of
70-mm film type 2403 and 2484 costs approximately $12 and $18, respec-
tively. Two cameras will be loaded with each type of film for optimum
multiband imagery. Since each camera holds 15 ft lengths, the film
expense per flight calculates to only $6.00. Under this contract, one
hour of flight time, required to exhaust the camera magazines, costs
approximately $115. Thus, a typical multiband flight would cost
approximately $221.
If the base and multiband flights are flown sequentially, the total cost
is found to be slightly higher. Black-and-white 9-in. film for the Zeiss
camera costs approximately $43. The multiband film for three cameras
containing film types 2403 and 2424 would cost $4.20. The 70-mm color
film costs approximately $18 for a 15 ft length. Since two flight costs
are required, the flight cost increases to $230. Therefore, the total
cost of the sequential flight would be $295.20. The above calculations
are independent of the initial camera costs. The four Hasselblad cameras
and accessories cost approximately $4500. The cost of the Zeiss camera,
including view finder, mounts, and other accessories, is estimated at
$28,000.
83
-------
SECTION IX
ACKNOWLEDGEMENTS
The work herein reported was performed by personnel of the Reconnaissance
Laboratory at McDonnell Aircraft Company. Mr. Charles L. Rudder was the
Principal Investigator on the Program. Mr. Charles J. Reinheimer
assisted in guiding the flight test effort and performed data correlation
and evaluation. Mr. Joseph L. Berrey participated in the flight program,
ground truth collection and image evaluation and mensuration. Messrs.
Raymond M. Bradley, William A. Dalton, Erich D. Kassler, John T. Smith,
and Robert E. Thompson each made invaluable contributions.
Messrs. James W. Walker and William P. Charbonier II of Surdex Corpora-
tion are gratefully acknowledged for their excellent performance in
flying the cartographic and multiband photography missions.
The support of the Agricultural and Marine Pollution Control Section,
Office of Research and Monitoring, Environmental Protection Agency and,
in particular, the direction provided by Mr. John Riley, the Project
Officer, are acknowledged with sincere thanks.
85
-------
SECTION X
APPENDICES
Page No.
A. Film and Filter Spectral Characteristics 88
Figure 1: Film Spectral Sensitivity 89
Figure 2: Film Spectral Sensitivity 90
Figure 3: Filter Transmission 91
Figure 4: Filter Transmission 92
Figure 5: Filter Transmission 93
Figure 6: Filter Transmission 94
Figure 7: Filter 25 Transmission 95
B. Photographic Processing 96
Table 1: Chemistry for Processing Color Film Type 2443
and 2448 97
Table 2: Chemistry for Processing of Film Types 2403,
2424 and 2475 98
Table 3: Environmental Conditions for Film Storage and
Handling 98
Figure 1: Typical H & D Curve for Film Type 2448 . . . 100
Figure 2: Typical H & D Curve for Film Type 2443 . . . 101
Figure 3: Gamma Variation with Film Processing .... 102
Figure 4: Typical H & D Curve for Film Type 2403 ... 103
Figure 5: Typical H & D Curve for Film Type 2424 . . . 104
Figure 6: Typical H & D Curve for Film Type 2475 . . . 105
Table 4: Sensitrometric Data 106
C. Steel Mill Waste Lagoon Multiband Photography 107
Figure 1: Steel Mill Waste Lagoon 108
Figure 2: Steel Mill Waste Lagoon 109
D. Oil Waste Dump 110
Figure 1: Baseline Imagery of Oil Refinery Waste Dump . 110
Figure 2: Oil Waste Dump 112
87
-------
APPENDIX A
FILM AND FILTER SPECTRAL CHARACTERISTICS
The identification of target materials from their reflective character^
istics can be achieved in aerial photography. The technique, called
multiband photography, requires recording spectral characteristics of
targets in multiple but separate images. Data recorded in the imagery
are determined by the combined film and filter spectral response.
Approximate target material identification is accomplished by inter-
preting the contrast, or lack of contrast, between the target and its
surround in the imagery. Oil products are known to reflect strongly in
the ultraviolet, blue, and red portions of the spectrum. Thus by
selecting the film and filter combinations that record only these por-
tions of the spectrum, one can enhance the contrast of oil products in
the photographic image. This technique is called positive contrast
enhancement. By choosing a filter that absorbs the reflected target
radiation, image contrast enhancement can also be achieved. This procea
is known as negative contrast enhancement. Filters have been chosen for-
multiband photographic evaluation that both transmit and absorb radia-
tion reflected from oil products. Additional filters were chosen that
transmitted adjacent and overlapping spectral bands in the visible port!
of the spectrum.
The spectral characteristics of Kodak Tri-X Aerographic film type 2403
Kodak Infrared Aerographic film type 2424, and Kodak Recording film
type 2475 are shown in Fig. 1. Only the film type number rather than th
descriptive title are specified in this report. Also, film and filter
combinations are specified as film/filter. Hence, the use of film type
2403 and filter 32 is denoted 2403/32. Film type 2475 is an instrument
recording film that was added to the program because of its increased
ultraviolet sensitivity. These films were used with various filters for
the multiband photography. The spectral characteristics of Kodak Ekta-
chrome MS Aerographic film type 2448 and Kodak Aerochrome film type 2443
are shown in Fig. 2. Film type 2448 is normal color film while film type
2443 is called false color film. These films contain three different
emulsions that are combined to provide the color imagery. For the
altitudes flown, film type 2448 was used with Kodak HF3 haze filter while
film type 2443 was used with a Kodak 12 filter. As recommended by
the manufacturer, a CC10M filter was also used with this particular batch
of film type 2443 to achieve the proper color balance. These filter
transmission curves are shown in Fig. 3a, 3b, and 3c, respectively.
The Kodak filters chosen for the multiband work were filters 18A, 47B,
39, 32, 35, 65, 75, 98, 99, and 25. The transmissions of these filters
are shown in Figs. 4, 5, 6 and 7. Filter 25 is the standard filter used
with film type 2424. Filters 18A, 47B, and 39 were chosen for their
ultraviolet and blue transmittance where oil is known to reflect strongly
These filters, when combined with film types 2403 and 2475, would restricl
the imaged radiation to these spectral bands. This can be seen by
88
-------
comparing Fig. 1 and Fig. 4. When these filters were combined with film
type 2424, the near-infrared portion of the spectrum (700 to 850 nm)
reflected from the target was also included.
Filters 99, 98, and 75 were chosen because of their absorption of blue
and ultraviolet radiation. Thus, the use of these filters with film
type 2403 and 2475 should yield images of oil products with negative
contrast enhancement. When these filters were combined with film type
2424, only the near-infrared and yellow-green portions of the spectrum
reflected from the target would be imaged. Oil has very little yellow-
green reflectance properties. A comparison of the imagery on film types
2403 and 2424 allowed us to determine the relative reflective properties
of oil in the near-infrared spectrum. Through the proper film and fil-
ter combination, the spectral characteristics of the target can be
determined.
Some oil products are known to reflect red as well as ultraviolet and
blue radiation. Filters 32 and 35 were included to evaluate the
addition of the red spectral region for the detection of oil products.
Film type 2475
Film type 2403
Film type 2424
Density = 0.3 above gross fog
250 300 350 400 450 500 550 600 650 700 750 800 850 900 950
Wavelength (nanometers)
Reproduced with permission from a
copyrighted Kodak publication
Figure 1 Film spectral sensitivity
GP71 1642-1
89
-------
c
01
GO
400 450
500
550 600 650 700
Wavelength (nanometers)
a) Film type 2443
750
800
90Q
1.0
250 300 350 400 450 500 550 600
Wavelength (nanometers)
b) Film type 2448
Reproduced with permission from
a copyrighted Kodak publication
650 700 75Q
Figure 2 Film spectral sensitivity
GP71-1642-2
90
-------
0.1%
£ Q
E s
. -"
^
'
300
Reproduced with permission from
a copyrighted Kodak publication
400 500 600
Wavelength (nanometers)
c) CC10M
Figure 3 Filter transmission
700
GP71-1642-3
91
-------
0.1%
100%
0.1%
a) 18A
100%
0.1%
b)47B
100%
200 300
Reproduced with permission from
a copyrighted Kodak publication
400 500 600 700
Wavelength (nanometers)
c)39
Figure 4 Filter transmission
800 900
GP71-1642-4
92
-------
0.1%
100%
0.1%
a) 32
10%
100% 0
b)35
0.1% 3
100%
200 300 400 500 600 700
Wavelength (nanometers)
c)65
Reproduced with permission from
a copyrighted Kodak publication
Figure 5 Filter transmission
800 900
GP71-1642-63
93
-------
0.1%
1% 2
in
C
I
10%
100% 0
a) 75
0.1%
c
to
1%
10%
100%
b)98
0.1%
100%
0
200
300 400 500 600 700
Wavelength (nanometers)
c)99
800 900
Reproduced with permission from
a copyrighted Kodak publication
Figure 6 Filter transmission
GP71-1642-62
94
-------
0.1%
100%
300
400 500 600 700
Wavelength (nanometers)
800 900
Reproduced with permission from
a copyrighted Kodak publication
Figure 7 Filter 25 Transmission
GP71-1642-61
95
-------
APPENDIX B
PHOTOGRAPHIC PROCESSING
Precision photographic processing of the 70-mm film is necessary for the
extraction of scientific data from multiband photographic images.
Precision processing insures the tonal variation in the multiband
images is a result of the target's reflectivity and not the result of
inconsistencies in the processing techniques. The processing equipment
should be lightweight and portable, capable of processing both black-
and-white and color films, have minimal operating procedures, and cost
less than a fully automatic processing system. This led to the
selection of a Nikor 70-mm Reel and Tank Processor. The reels had a
15 ft capacity and the tanks had a 1/2 gal. capacity. The Nikor Film
Loading Device was also selected. Film drying was accomplished with an
Oscar Fisher forced hot air cabinet dryer. Agitation for black-and-white
film processing was mechanically applied by a Pierce "ROK-IT" Agitator.
The color processing chemistries considered were Ektachrome Process E-3
and E-4. Both of these produce nearly the same results when used to
process color films types 2448 and 2443. Since Process E-4 has two
more solutions than Process E-3, E-3 processing was selected. The E-3
processing steps for color films types 2443 and 2448 are shown in
Table 1.
The selection of the chemistry for processing of the three black-and-
white film types 2403, 2424, and 2475 was more complicated. The design
goal here was a single chemistry, common to the three films, which
would produce the desired contrast range and film speed. Sensitometric
experiments were conducted and analysis of the derived data indicated
that development in D-19 would satisfy the requirements. The D-19
chemistry and functions relating to the processing of black-and-white
films type 2403, 2424, and 2475 are contained in Table 2.
Film Storage and Handling
It is a well recognized fact that unprocessed photographic film is
perishable. Its sensitometric properties will deteriorate slowly with
time. The deterioration is accelerated by high relative humidities and
high temperatures. Sensitometric degradation is usually reflected in a
speed loss, an increase in the base fog level, a contrast change, or
any combination thereof. Color films can also exhibit changes in color
balance. These known sensitometric changes in photographic films will
not be encountered if film is stored and handled under proper conditions
for reasonable periods of time. Films used for the airborne data col-
lection phase of this contract were given strict environmental
protection, from shipment through the processing. Films were packed in
dry ice and shipped via air by the manufacturer. Upon receipt, the
film was immediately stored in a deep freeze (long term) or in a refrig-
erator (short term). The film, after removal from refrigeration, was
96
-------
Table 1 Chemistry for processing color film types 2443 and 2448
Step
1
2
3
4
5
6
7
8
9
10
11
12
13
14
Function
Develop
Rinse
Harden
Wash
Re-expose
Develop
Wash
Clear
Bleach
Rinse
Fix
Wash
Stabilize
Dry
Kodak Ektachrome
process E-3
First developer
Water (4 gpm)
Hardener
Water (4 gpm)
No. 2 photoflood
Color developer
Water (4 gpm)
Clearing bath
Bleach
Water (4 gpm)
Fixer
Water (4 gpm)
Stabilizer
Forced hot air
Time
(min)
10
1
3
3
1/2
15
5
5
8
1
6
8
1
20
Temperature
(°F)
75 + 1/2
75+2
75 + 1/2
75+2
/5+2
75 + 1/2
75+2
75 + 1/2
75 + 1/2
75+2
75 + 1/2
75+2
75+2
110+8
GP71-1642-65
tempered at ambient room temperature (73°F) for not less than eight
hours prior to the actual camera loading. Upon completion of magazine
loadings, all unused film was resealed in metal cans and stored at
ambient room temperature. All film handling (except that which
occurred during the airborne flights) was conducted under ambient room
temperatures of 73°F, or 75°F in the case of film processing. Actual
environmental conditions relative to storage and handling of the films
are shown in Table 3. Analysis of sensitometric data showed that
neither a loss of film speed, an increase in fog level, or a change in
color balance had occurred in any of the processed films. These
results are attributed to the careful handling and storage of the
unprocessed film as well as to precision processing techniques. Because
of the interdependence of the various parameters, such as temperature
humidity and the time the film is maintained at unfavorable conditions,
it is difficult to estimate the amount of film degradation if these
procedures are not followed.
97
-------
Table 2 Chemistry for processing of film types 2403, 2424 and 2475
Step
1
2
3
4
5
6
7
8
Function
Develop
Stop
Fix
Rinse
Clearing
Wash
Wetting agent
Dry
Agent
Kodak D-19
Kodak Indicator Stop
Kodak Rapid Fixer
Water (4 gpm)
Kodak Clearing Agent
Water (4 gpm)
Kodak Photo-Flo 200
Forced hot air
Time
(min)
8
1
5
1
2
5
1
15
Temperature
(°F)
75 + 1/2
75 + 1/2
75 + 1/2
75+2
75 + 1/2
75+2
75+2
160 + 10
GP71-1642-66
Table 3 Enviornmental conditions for film storage and handling
Type of film storage
and handling
Long term
Short term
Tempering
Loading
Processing
Environmental conditions
Type area
Deep freeze
Refrigeration
Air conditioned
Air conditioned
Air conditioned
Temperature
-4!"-4
40+3
73 + 1
73 + 1
75 + 1
Relative
humidity
70+6
44+6
45+5
45+5
45+5
GP71-1642-67
98
-------
Film Processing Procedure
For color processing the solution tanks were partially immersed in re-
circulating water temperature controlled at 75°F. Agitation was
performed by the tip-and-tilt method as prescribed for Nikor Reel and
Tank Processing. As described earlier, the chemistry used was Kodak
Ektachrome Process E-3, 1/2 gal. size. To minimize the chances of
chemical contamination, a replenishment system was not used. Instead,
freshly mixed chemicals were used for each 15 ft of film processed.
Sensitometric control step wedges were processed with each 15 ft roll,
and the derived characteristic H&D curves exhibited good process control
and good color balance. Typical characteristic H&D curves for color
film types 2448 and 2443 are shown in Fig. 1 and 2, respectively.
Black-and-white film processing was also accomplished by using Nikor
stainless steel 70-mm by 15 ft reels and 1/2 gal. tanks. Solution tanks
(except the developer) were partially immersed in temperature-controlled
(75°F), circulating water. The tank with the developer solution sat on
a Pierce Company, "ROK-IT" Agitator. The "ROK-IT" is an electro-
mechanical device that rocks the developer solution. Automatic, rather
than manual, agitation was used to provide repeatable processing
conditions.
Gamma (y) aim points, with allowable upper and lower limits, were estab-
lished for each type of film. The selected gamma aim points were those
that would provide the desired emulsion speed and contrast range.
Sensitometric data derived from step wedges processed with the aerial
imagery showed that in all cases, processing control was within the
established limits. (See Fig. 3.) Process control step wedges were
also processed with each reel of film. Densities of the wedges were
measured and their numerical values plotted against the log of exposure.
Other Sensitometric data and typical characteristic H&D curves relative
to the processing of the three films, types 2403, 2424, and 2475, are
shown in Figs. 4, 5, 6, respectively. From the resultant characteristic
curves, other necessary data such as minimum and maximum density, gamma
and emulsion speed were obtained. Averages of these data are shown in
Table 4.
99
-------
Sensitometric Data
Step tablet-Kodak 007 ST403; Sensitometer illumination - tungsten; Exposure - tungsten; Exposure time -
.2 sec; Meter candle seconds - 340; Sensitometer filtration - Kodak # 309, Wratten #96, neutral
density - 1.4; Development method - reel & tank; Developer - Ektachrome E-3; First development time -
10 min; Solution temperature - 75°F; Agitation method - manual tip and tilt; Densitometer filtration -
Wratten #92, 98, & 99; Red j- 2.9, green y - 1.79 & blue 7 - 1.64; Emulsion speed (AEI) - red 9, green 3
& blue 1; Note - emulsion speed is effective for indicated processing only.
c
0)
Q
Figure 1 Typical H&D curve for film type 2448
GP71-1642-7O
100
-------
Sensitometric Data
Step tablet-Kodak 007ST403; Sensitometer illumination - tungsten; Exposure time - .2 sec; Meter candle
seconds - 340; Sensitometer filtration - Wratten #12 & 96, Kodak #301 & 309, neutral density - 1.4;
Development method - reel & tank; Developer - Ektachrome E-3; First development time - 10 min;
Solution temperature - 75°F; Agitation method - manual tip and tilt; Densitometer filtration - Wratten
#92, 98 & 99; Infrared y - 4.22, green 7 - 3.07, blue y - 5.45; Emulsion speed (AEI) - infrared 1.4, green-
12.5, blue - 17; Note - speed is effective for indicated processing only.
0.0
0.0
Figure 2 Typical H&D curve for film type 2443
GP71-1642-69
101
-------
Control wedge no.
Upper limit
7 aim point
Lower limit
+0.15
+0.10
+0.05
1.51
-0.05
-0.10
-0.15
Film type 2403
Control wedge no.
Upper limit +0.15
+0.10
+0.05
7 aim point 1.82
-0.05
-0.10
Lower limit -0.15
1
/
/
2
A
3
v
V
4
S*^
5
X
6
X.
•*•«»
7
^
^^
8
/
9
/
Film type 2424
Control wedge no.
Upper limit +0.15
+0.10
+0.05
7 aim point 1.51
-0.05
-0.10
Lower limit -0.15
1
\
s
2
V
3
*.
/
w
4
X.
^•v
5
•— «
Film type 2475
Figure 3 Gamma variation with film processing
GP71-1642-64
102
-------
Sensitometric Data
Step tablet-Kodak 007ST403; Sensitometer illumination - tungsten; Exposure time - .2 sec; Meter candle
seconds - 340; Sensitometer filtration - Wratten #96, neutral density - 3.1; Development method - reel
& tank; Developer - D-19; Development time - 8 min; Solution temperature - 75°F; Agitation method -
mechanical rocking; j - 1.51; Emulsion speed (AEI) - 284; Note -speed is effective for indicated pro-
cessing only.
0.0
Log Exposure
Figure 4 Typical H&D curve for film type 2403
GP71-1642-72
103
-------
Sensitometric Data
Step tablet-Kodak 007ST403; Sensitometer illumination - tungsten; Exposure time - .2 sec; Meter candle
seconds - 340; Sensitometer filtration - Wratten #96, neutral density - 2.78; Development method - reel
and tank; Developer - D-19; Development time - 8 min; Solution temperature - 75°F; Agitation method -
mechanical rocking; j - 1.72; Emulsion speed (AEI) - 126; Note - emulsion speed is effective for indicated
processing only.
3.0
2.0
c
o>
Q
1.0
0.0
3.0
2.0 1.0
Log Exposure
Figure 5 Typical H&D curve for film type 2424
GP71-1642-71
104
-------
Sensitometric Data
Step tablet-Kodak 007ST403; Sensitometer illumination - tungsten; Exposure time - .2 sec; Meter candle
seconds - 340; Sensitometer filtration - Wratten #96, neutral density - 3.2; Development method - reel
& tank; Developer - D-19; Development time - 8 min; Solution temperature - 75°F; Agitation method -
mechanical rocking; j - 1.51; Emulsion speed - AEI 330 & ASA 900; Note - emulsion speed is effective
for indicated processing only.
3.0
2.0
c
o>
Q
§
1.0
0.0
3.0 2.0
Log Exposure
1.0
Figure 6 Typical H&D curve for film type 2475
GP71-1642-68
105
-------
Table 4 Sensitrometric data
Film
type
2403
2424
2443
2448
2475
Emulsion
Black and white
(panchromatic)
Black and white
(panchromatic)
Color (infrared) Red
Green
Blue
Color (reversal) Red
Green
Blue
Black and white
(panchromatic)
Average
Dmin
.09
.16
.27
.12
.15
.07
.12
.17
.32
"max
2.71
2.49
3.57
3.44
>4.00
2.66
2.80
3.03
2.41
Gamma
1.51
1.71
4.25
3.03
5.12
2.12
1.82
1.71
1.51
Effective
film speed
AEI 284
AEI 126
AEI 1.4
AEI 12.5
AEI 17
AEI 9
AEI 3
AEI 1
ASA 900
GP71-1642-76
106
-------
APPENDIX C
STEEL MILL WASTE LAGOON MULTIBAND PHOTOGRAPHY
In the steel industry, waste storage lagoons contain a variety of mater-
ials. These materials include lubricating oil from the rolling mill
cooling system, rust or iron which is acidicly removed from stored steel
materials, and water from various cooling towers. The cooling water
contains chemicals that are used to eliminate scale and fungus from the
cooling towers. The lagoon that contains the above materials is treated
with caustic soda to neutralize the acid and settle out solids (iron
scale) and is then passed through a filtration pond before it is emptied
into the river.
The steel mill waste lagoons shown in the color aerial photograph in
Fig. la are spectrally characterized by a variety of colors. Before
being treated with caustic soda, the lagoons are yellowish red in color.
After treatment, the ponds acquired a rust-colored or a black appearance.
The rust-colored area is iron or scale settlement. A ground truth team
learned the black color is a residue resulting from the reaction of
spent caustic soda with water. A photograph of a typical pond obtained
during an initial ground truth mission is shown in Fig. Ib. Note the
rust color appearance of the pond in the upper portion of the photograph
while the pond to the right appears free of pollutants. The black
and rust-colored appearance of the ponds settling area is contrasted
with the untreated yellow portion of the lagoon labeled A.
The value of multiband photography is in the identification of the mater-
ials present in these lagoons. These results can be applied to areas
located adjacent to inland waterways to identify unknown spills and thus
located potential and real spill sources threatening inland waterways.
The multiband flights revealed that the film/filter combinations 2403/99,
2403/65, 2403/32, 2424/99, and 2424/35 gave the best image contrast for
this area. These film/filter combinations are shown in Figs. Ic and Id
and 2a, 2b, and 2c, respectively. Only small differences were found in
the imagery recorded with film type 2403 and filters 99, 98 and 75. The
same conclusion was derived for these filters coupled with film type
2424. Similarly, there were only slight differences in the imagery
recorded with filters 32 and 35. As can be seen from Figs. 1 and 2, the
oil water area imagery is slightly darker for the 99 filter coupled with
both film types 2403 and 2424. The imagery recorded with film/filter
combination 2424/35 gives better area definition than the 2424/99
imagery. This is indicative of a spectral return in the ultraviolet and
blue portions of the spectrum. The imagery recorded with film/filter
combination 2403/32 and 2403/65 even better defines this area because of
the water-reflected visible radiation. The film/filter combination
2424/35 imagery shows strong water absorption in the near-infrared.
Thus it is difficult to distinguish oil from the water.
The yellowish lagoon has a light tone in the film/filter combination
2403/32 and 2403/65 imagery. On the imagery recorded with film/filter
107
-------
t'/l
ttl/lr i
b) Ground truth photograph
a) 2448/HF3
c) 2403/99
d) 2403/65
Figure 1 Steel mill waste lagoon
combinations 2403/99, 2424/99, and 2424/35, this area is dark in tonal
quality, representing a complete contrast reversal. This area strongly
reflects ultraviolet, blue, and green visible radiation and absorbs
near-infrared radiation.
The rust-colored area shows small contrast variations in the imagery
recorded on five film/filter combinations shown in Figs. 1 and 2. The
imagery of this area obtained with film/filter combinations 2403/32 and
2424/35 is lighter than that obtained with filter 99 and film types 2403
and 2424. This is anticipated since filters 32 and 35 transmit the red
portion of the spectrum. The imagery from film/filter combinations
2403/65 is darker in this area than that taken with film/filter combina-
tion 2403/32 because of blockage of the red spectral region. The
imagery of this area taken with 2403/99 is lighter than that obtained
108
-------
a) 2403/32
b) 2424/99
c) 2424/35
Figure 2 Steel mill waste lagoon
from 2403/65. This indicates some reflectance of this area in the 500
to 600 nm range. The imagery obtained with film types 2424/35 and
2424 99 is found to be darker than that obtained with 2 03 2 and
2403/99 respectively. This indicates some near-infrared absorption by
this area. A comparison of the imagery obtained with film/filter
combinations 2424/35 and 2424/99 indicates that the red reflectance is
greater than the near infrared absorption. Thus, the various film/fliter
combinations allow ,the spectral characteristics of the materials
under investigation to be determined. One film/filter combination can-
not be used effectively for detecting the three different areas defined
in this lagoon. Color photography reveals the maximum amount of
information.
109
-------
APPENDIX D
OIL WASTE DUMP
Another oil waste area of 63 acres was covered during this project and is
shown in Fig. 1. It is located 7,000 ft from the river and does not repre-
sent a direct threat to the inland waterway. In addition, the area is
located on a flood plain and no distinct drainage pattern is observed.
The area is of interest for two reasons, First, it represents a waste
area that differs distinctly from other oil waste areas already discussed.
The oil waste in this area is located in small patches dispersed among
small, stagnant water ponds, vegetation, filter cakes and emulsified oil
waste, which accounts for the mottled appearance of the image shown in
Fig. 1. Second, the results of the multiband image analysis can
in similar areas located adjacent to inland waterways to identify unknown
spills. Upon identification, the potential or real threat of the spill
source to the inland waterway can be evaluated.
Figure 1 Baseline imagery of oil refinery waste dump
110
-------
An aerial color photograph of the area obtained on the multiband flights
is shown in Fig. 2a. An accompanying ground truth photograph of the oil
lying on the ground is shown in Fig. 2b. The multiband imagery revealed
the best film/filter combinations for contrast enhancement were 2403/32,
2403/99, 2424/32, and 2424/99. Images of the area using these film/filter
combinations are shown in Figs. 2c, 2d, 2e and 2f, respectively. As con-
cluded in the preceding section, color photography appears best in iden-
tifying the emulsified oil waste and filter cakes. The emphasis here is
on separating the black oil waste from the background material. The
imagery recorded with filter 99 for both film types 2424 and 2403 gives
the maximum negative enhancement of the oil to the background when com-
pared with the imagery recorded with filter 32. Filter 99's absorption
of the ultraviolet, blue, and red radiation reflected by oil is respon-
sible for this. When comparing film type 2424 and 2403 imagery, better
oil background contrast with type 2424 is observed because of the near-
infrared reflectance of the background. Because of the water near-infrared
absorption, however, it is sometimes difficult to distinguish oil from
water on film type 2424 imagery. Furthermore, the moisture content of
various ground regions produces a variety of shades of gray on film type
2424, tending to distract the photointerpreter. On film type 2403 imagery,
these shades of gray are subdued as the water ponds do not absorb as much
visible radiation making it easier to detect oil from water. For these
reasons fxlm/filter combinations 2403/99 and 2403/32 would be more effec-
f area ^ " detect^g oil waste in this particular
or arcd. ^
111
-------
b) Ground truth photograph
a) 2448/HF3
c) 2403/32
d) 2403/99
e) 2424/32
f) 2424/99
Figure 2 Oil waste dump
GP71-1642-37
112
-------
~ 1 .Sn/>/'-< I l-'n-lil it Group
05B
SELECTED WATER RESOURCES ABSTRACTS
INPUT TRANSACTION FORM
Reconnaissance Laboratory, McDonnell Aircraft Company
McDonnell Douglas Corporation, Box 516, St. Louis, Missouri 63166
AERIAL SURVEILLANCE SPILL PREVENTION SYSTEM
J Q Aulhorfs)
C. L. Rudder
C. J. Reinheimer
J. L. Berrey
16 15080 HOK EPA, ORM Contract No. 68-01-0140
21 »
22
Citation
Environmental Protection Agency report
number EPA-R2-72-007, August 1972.
23
Descriptors (Starred first)
Water Pollution Sources, Oil, Chemicals, Remote Sensing, Aerial Sensing,
Photography, Photogrammetry
or \ Identifiers (Starred first)
—-I Multiband Photography, *Photographic Mensuration, *Color Photography, Photointer-
pretation, Hazardous Materials
27
A bstract
An aerial surveillance system, consisting of four Hasselblad cameras and a Zeiss RMK
1523 camera, was evaluated for the remote detection of both real and potential spills
threatening inland waterways. Twenty-three multiband and baseline missions were
flown over oil refineries and other industrial sites located adjacent to the
Mississippi River. Baseline flights were effective in counting storage tanks, loca-
ting and identifying storage equipment and pipeline systems and determining dike con-
ditions. Stereoscopic analysis of baseline imagery was used to estimate the height
of tanks and dikes, drainage patterns and the area of openly stored waste products.
The multiband imagery was obtained by combining each of nine filters with each of
three different black and white films. Spectral contrast image enhancement was
accomplished by either suppressing or transmitting the target reflected radiation
through proper film/filter selections. Spills, effluents and waste areas were hence
identified on the multiband imagery. Normal and false color imagery was evaluated
with the multiband imagery to determine the best film/filter combinations for the
areas of interest. Finally the personnel, equipment and procedures required to
implement an aerial surveillance spill prevention system were determined.
/I6s'rac"t;harles J. Reinheimer
Institution
McDonnell Aircraft Company
*R 102 (RE
WWSI C
U. S. GOVKHNMKNT I'HINTINC. OKFICK : 1"72 C) - 4H6-910
------- |