THE EFFECTS OF PHOTOGRAPHIC GROUND RESOLUTION ON PHOTO INTERPRETATION

• Document Type: CREST
• Collection: NGA Records (Formerly NIMA)
• Document Number (FOIA) /ESDN (CREST): CIA-RDP79B00873A001600040034-7
• Release Decision: RIPPUB
• Original Classification: S
• Document Page Count: 18
• Document Creation Date: December 28, 2016
• Document Release Date: August 30, 2012
• Sequence Number: 34
• Publication Date: July 1, 1970
• Content Type: REPORT

Declassified in Part - Sanitized Copy Approved for Release 2012/08/30: CIA-RDP79B00873AO01600040034-7
L/ U
SECRET
B
a
0
0
8
a
a
0
a
0
a
THE EFFECTS OF PHOTOGRAPHIC
GROUND RESOLUTION ON
PHOTO INTERPRETATION
FINAL REPORT
SECRET
Declassified in Part - Sanitized Copy Approved for Release 2012/08/30: CIA-RDP79B00873AO01600040034-7
 L Declassified in Part - Sanitized Copy Approved for Release 2012/08/30: CIA-RDP79B00873AO01600040034-7
SECRET
B
B
a
8
a
a
a
B
a
a
0
a
r,
a
a
a
THE EFFECTS :OF PHOTOGRAPHIC
GROUND RESOLUTION ON
PHOTO INTERPRETATION
FINAL REPORT
SECRET
a
Declassified in Part - Sanitized Copy Approved for Release 2012/08/30: CIA-RDP79B00873AO01600040034-7
 Declassified in Part - Sanitized Copy Approved for Release 2012/08/30: CIA-RDP79B00873A001600040034-7
SECRET
U
r
THE EFFECTS OF PHOTOGRAPHIC
GROUND RESOLUTION ON PHOTO INTERPRETATION
Final Report
E
July 1970
SECRET
Declassified in Part - Sanitized Copy Approved for Release 2012/08/30: CIA-RDP79B00873A001600040034-7
 C
Declassified in Part - Sanitized Copy Approved for Release 2012/08/30: CIA-RDP79B00873A001600040034-7
SECRET
Page
C
C
C
WORK STATEMENT . . . . . . . . . . . . . . . . . .. 2
INTRODUCTION . . . . . . . . . . .
Experimental Photographs . . . . . . . . . . 8
Targets . . . . . . . . . . . . . . . . . . . . . . 9
Experimental Design. . . . ... . . . . . . . . . . 10
Subjects . . . . . . . . . . . . . . . . . . . . 11
Data Collection . . . . . . . . . . . . . . . . . . 11
Data Analysis . . . . . . . . . . . . . . . . . . . . 11
Technical Report . . . . . ... . . . . . . . . . . . 12
WORK PERFORMED . . . . . . . . . . . . . . . . . . . . . .
STUDY PLAN . . . . . . . . . . . . . . . . . . . . .
I iii
SECRET
Declassified in Part - Sanitized Copy Approved for Release 2012/08/30: CIA-RDP79B00873A001600040034-7
 E1
Declassified in Part - Sanitized Copy Approved for Release 2012/08/30: CIA-RDP79B00873AO01600040034-7
SECRET
U
LIST OF FIGURES
Figure
Li,
Page
1 Second generation paper prints of the Aber-
deen Proving Ground tank and artillery
museum . . . . . . . . . . . . . . . . . . . . 5
c
C
n
Second generation paper prints of targets at
Ft. Meade . . . . . . . . . . . . . . . . . 7
iv
SECRET
Declassified in Part - Sanitized Copy Approved for Release 2012/08/30: CIA-RDP79B00873AO01600040034-7
 Declassified in Part - Sanitized Copy Approved for Release 2012/08/30: CIA-RDP79B00873AO01600040034-7
L
SECRET
THE EFFECTS OF PHOTOGRAPHIC
GROUND RESOLUTION ON PHOTO INTERPRETATION
Final Report
C
This contract was a continuation of an effort to deter-
mine the effects of photographic ground. resolution on the
accuracy of the information produced by photo interpreters
(PIs). Three comprehensive studies had been done under a
previous contract In. the. first,
the targets were U. S. aircraft;. in the second, they were.
domestic ICBM sites; and in the third, they were scale models
of mobile foreign radars. Technical reports describing these
studies have been submitted to the sponsor.
The studies showed clearly that the effects of ground
resolution on photographic interpretation are target specific;
one class-of targets requires better resolution than another
class for complete exploitation.
It had been suggested that ground forces as a class of
targets might require even better ground resolutions than
radars for the complete exploitation of photography. There-
fore, it was proposed that additional studies similar to the.
aircraft, missile, and radar studies be conducted with ground
equipment as the targets; more specifically,. it was proposed
that two studies be done in the following two phases:
c
Phase I - Model Study and Preparation for the Field
Study
1a: A study of the effects of ground resolution on
the identification of military vehicles using
SECRET
Declassified in Part - Sanitized Copy Approved for Release 2012/08/30: CIA-RDP79B00873AO01600040034-7
 Declassified in Part - Sanitized Copy Approved for Release 2012/08/30: CIA-RDP79B00873A001600040034-7
SECRET
scale models as targets (the "model" study).
1b: The preparation of materials for a study
of the effects of ground resolution on the
interpretation of ground forces using real ground
force equipment and personnel as targets (the
"field" study).
Phase II - Execution of the Field Study
The proposal, which was to conduct Phase I only, was
supported and a new contract was written. The starting date
was 5 June 1969, and the termination date (after an extension
in time without additional funds) was 15 July 1970.
WORK STATEMENT
In Phase Ia, the model study, the following tasks were
to have been performed:
Task 1. Preparation of photographic stimulus
materials.
Task 2. Development of the experimental design,
instructions, and PI response recording
method.
Task 3. Selection of experimental subjects and
collection of experimental data.
Task 4. Analysis of data.
Task 5. Preparation of the technical report and
briefings.
Li
In Phase Ib, preparation for the field study, the
following tasks were to have been performed:
Task 1. Selection and procurement of the required
aircraft and camera.
Task 2. Selection, procurement, and placement of
appropriate targets.
Task 3. Construction or procurement and placement
of test targets.
2
SECRET
Declassified in Part - Sanitized Copy Approved for Release 2012/08/30: CIA-RDP79B00873A001600040034-7
 Declassified in Part - Sanitized Copy Approved for Release 2012/08/30: CIA-RDP79B00873A001600040034-7
SECRET
Task 4. Collection of "ground truth" information.
Task 5. Collection of photography.
J
E
0
F.
The model study, Phase Ia, was completed, several brief-
ings were given in Washington, and the.technical report was
published in October'1969. The work is described in:
The model study was the first effort to determine the
relation between line-scan and photographic images in terms
of photo interpreter target identification performance. Two
line-scan image variables were investigated: signal-to-noise
ratio and number of scans per target. One photographic image
variable was investigated: ground resolution. The targets
were models of tanks and miscellaneous vehicles, and the
subjects were 50 professional photo interpreters. The data
were analyzed to determine what combinations of the two line-
scan variables were equivalent to photographic ground resolu-
tion in terms of percent correct target identification.
The results with the two types of imagery and the'rela-
tions between the results should be valuable to system de-
signers. However, additional research on line-scan imagery
should be done.
Phase lb was also completed. The photographic acquisi-
tion requirements established jointly by the sponsor,. Auto-
metric/Raytheon, and[~were:
1. One-inch ground resolution as determined from
one cycle on a bar target.
2. Stereo availability; 20%.convergence angle or
60%.overlap.
. 3
SECRET
Declassified in Part - Sanitized Copy Approved for Release 2012/08/30: CIA-RDP79B00873A001600040034-7
 Declassified in Part - Sanitized Copy Approved for Release 2012/08/30: CIA-RDP79B00873A001600040034-7
rl
SECRET
3. A frame camera flown in a vertical position;
conventional black-and-white film.
4. Little haze, minimum cloud cover, and a sum
angle of about 60? above the horizon..
5. Original negative processed with a density
wedge.
Ell
An investigation was made of the camera/vehicle sys-
tems that could fulfill the requirements. Few were in exis-
tence but an adequate one was obtained through the sponsor.
Arrangements were made with Aberdeen Proving Ground to
photograph the vehicles and weapons in the museum and other
ground order-of-battle targets. Also, permission was ob-
tained to place test targets on the ground to obtain "ground
truth" data, photographs, and measurements. These tasks. were
performed.
Several flights were flown to determine the appropriate
altitude and camera settings. Many delays due to equipment
malfunctions, camera and processing problems,. and weather
were encountered.
On 17 September 1969, the acquisition flight was flown
over the museum at Aberdeen. An effort was made to get all
of the targets on one or two frames. From three-bar target
estimates, the ground resolution of the photography was
approximately 1.5".
Second generation paper prints of the Aberdeen museum
targets are shown in Figure 1.
At the time of the overhead flights, luminance measures
were made on the ground and ground photographs of the targets
were taken.
At the suggestion of the sponsor, additional flights were.
made over Ft. Meade, Maryland, to obtain a larger sample of
n
targets for the study. The ground resolution of the photog-
raphy obtained was estimated at 1.5" from three-bar targets.
F
4
SECRET
Declassified in Part - Sanitized Copy Approved for Release 2012/08/30: CIA-RDP79B00873A001600040034-7
 n Declassified in Part - Sanitized Copy Approved for Release 2012/08/30: CIA-RDP79B00873AO01600040034-7
fl
E
E
SECRET
Fig. 1. Second generation paper prints of the Aberdeen
Proving Ground tank and artillery museum.
F
S
-SECRET
Declassified in Part - Sanitized Copy Approved for Release 2012/08/30: CIA-RDP79B00873AO01600040034-7
 C
Declassified in Part - Sanitized Copy Approved for Release 2012/08/30: CIA-RDP79B00873A001600040034-7
SECRET
Second generation paper prints of the Ft. Meade photography
are shown in Figure 2. After the photography had been ob-
tained, it was necessary to perform additional tasks in
preparation for the experimental study:
1. Obtain photographic images representing
five ground resolutions and prepare an MTF
and ground resolution estimate for each.
2. Obtain "ground truth" for the additional
photographic coverage.
r
r
n
c
c
G
17
3. Prepare a study plan.
Item 1 above, obtaining photographs representing five
ground resolutions, implied obtaining five ground resolutions
of one or two scenes. Actually, six ground resolutions of
six scenes were prepared by Eastman Kodak from the original
negatives. The need.for additional "ground truth" informa-
tion was a result of obtaining the additional targets-and.
scenes at Ft. Meade. That "ground truth" was obtained.
A study plan was prepared but this work was done before
the photographs were prepared by Eastman Kodak. Thus, it
was not possible to perform a pilot study, and the plan can-
not be considered complete. It was not possible, for ex-
ample, to determine how difficult it is going to be to point
at the targets to be identified, how long it is going to take
the PIs to "read out" each frame, or the best way to index
and present the "ground truth" information.
It is assumed now that the method will be essentially
the same. as that used in the previous studies. Experienced
PIs will view the targets at each ground resolution, starting
with the poorest and ending with the best. At each resolu-
tion, they will attempt to identify each target by comparing
its image in the photographs with the "ground truth" information.
6
SECRET
`" Declassified in Part - Sanitized Copy Approved for Release 2012/08/30: CIA-RDP79B00873A001600040034-7
 0
0
0
0
a
0
0
0
0
Declassified in Part - Sanitized Copy Approved for Release 2012/08/30: CIA-RDP79B00873AO01600040034-7
SECRET
Fig. 2. Second gerierationpaper prints of targets at Ft.
Meade.
7
SECRET
Declassified in Part - Sanitized Copy Approved for Release 2012/08/30: CIA-RDP79B00873AO01600040034-7
 Declassified in Part - Sanitized Copy Approved for Release 2012/08/30: CIA-RDP79B00873AO01600040034-7
0
0
0
0
0
0
a
0
a
0
0
0
SECRET
F?ig. 2 (Cont.)
The data will be analyzed to determine the mean, standard
deviation, and standard error of the mean of percent correct
identifications at each ground resolution. Tests of the
statistical significance of differences in identification
performances between and among ground resolutions will be
made. A description of the method and the results and their
implications will be presented in a technical report.
Following is a more detailed description of the method.
Experimental Photographs.
The ground resolution of the photography is approxi-
mately) The original negatives were processed to obtain
8
aSECR
ET
Declassified in Part - Sanitized Copy Approved for Release 2012/08/30: CIA-RDP79B00873A001600040034-7
 Declassified in Part - Sanitized Copy Approved for Release 2012/08/30: CIA-RDP79B00873AO01600040034-7
SECRET
E
six ground resolutions and scales that are relevant to
existing and proposed systems, as shown in the table below.
Targets
As previously mentioned, representatives of the sponsor
and other intelligence organizations suggested that the effects
of ground resolution on the identification of ground order-
of-battle targets should be investigated. Consequently, a
study was done first of the effects of ground resolution on
the accuracy of identification of scale models of military
vehicles--tanks and other types of vehicles.
The. sponsor requested that an additional study with real
ground order-of-battle targets be conducted. So photographs
of the World War II tank/artillery museum at Aberdeen and
more modern targets at Ft..Meade were obtained.
There may be a question concerning whether or not the
targets are representative of the ground order-of-battle tar-
gets that PIs seek to identify today. They appear to be.
First, some of the World War II tanks and artillery pieces at
Aberdeen and the vehicles photographed at Ft. Meade are in
use today; second, the external configurations of the tanks
and artillery pieces developed since World War II, and being
used today, are not signifcantly different from those used in
9
SECRET
Declassified in Part - Sanitized Copy Approved for Release 2012/08/30: CIA-RDP79B00873AO01600040034-7
 Declassified in Part - Sanitized Copy Approved for Release 2012/08/30: CIA-RDP79B00873AO01600040034-7
SECRET
E
World War II; and, third, the visual discriminations required
to identify World War II ground order-of-battle targets are
not different from the visual discriminations required to
identify modern ground order-of-battle targets.
Experimental Design
The major question in designing a study of the effects
of ground resolution on PI performance is whether to use a
within- or between-subjects design. In the within-subjects
design, all subjects view the imagery at all ground resolu-
tions and thus serve as their own controls. The advantage
of such a design is that because the subjects serve as their
own controls, some experimental control is maintained over
relevant variables such as knowledge, experience, and apti-
tude. There are disadvantages to the within-subjects design.
The number of hours required per subject is much larger, and
subject motivation and performance can deteriorate over time.
There may be uncontrolled "order effects," and the order
effect of greatest concern is the learning that takes place
during the course of the experiment.
L In the between-subjects design,. a different group of
subjects views the imagery at each ground resolution. Thus,
when a comparison is made among the performances at the
various ground resolutions, not only are the effects of ground
resolution being compared, individual differences in knowledge,
experience, and aptitude are also being compared.
In the previous experiments, within-subjects designs
have been used. If the task is not too time consuming, the
same type of design will be used again. A pilot study will
be done to determine how long it takes a subject to complete
the entire target identification task working from the poorest
imagery to the best. If the time taken is considered reason-
able, a within-subjects design will be used. If the time re-
quired is considered too long, a combination within- and
10
SECRET
Declassified in Part - Sanitized Copy Approved for Release 2012/08/30: CIA-RDP79B00873AO01600040034-7
 E
Declassified in Part - Sanitized Copy Approved for Release 2012/08/30: CIA-RDP79B00873AO01600040034-7
SECRET
between-subjects design will be used, as illustrated below.
GROUND RESOLUTION
Group
I
Group
II
Group
. I
Comparisons betweei
Group
II
Group
I
Group
II
will be within subjects. Other comparisons will
E
J
n
be between subjects. Obviously, a pilot study must be done
before final design decisions can be made.
Subjects
Professional PIs with a minimum of one year of recent
operational experience will serve as subjects and an effort
will be made to use only ground order-of-battle specialists.
From 20 to 30 subjects should be enough to achieve adequate
reliability, yet not seriously interfere with operations.
The data collection will be conducted at NPIC. A room
with three viewing stations comparable to operational sta-
tions will be set up and an effort will be made to have three
subjects perform at a time. Each station will include a light
table, a B and L microscope, and a tube magnifier.
Data Analysis
Means, standard deviations, and standard errors of the
mean of percent correct target identifications at each ground
resolution will be computed. Different classes of targets
will be treated separately. Analyses of variance will be
performed to determine the overall significance level of
ground resolution as a function of.target type. Statistical
tests of the differences between pairs of means will be made
11
SECRET
L_- 4
Declassified in Part - Sanitized Copy Approved for Release 2012/08/30: CIA-RDP79B00873AO01600040034-7
 Declassified in Part - Sanitized Copy Approved for Release 2012/08/30: CIA-RDP79B00873AO01600040034-7
EI
r.
as required. Other analyses will be performed if they appear
possibly to reveal relevant information.
The data analysis will be done on the~BM 1130 com-
puter, using programs especially designed for the study.
Technical Report
The technical report will contain a summary, an intro-
duction,.a detailed description of the method and results,
conclusions, and discussion. Briefings will be given at the
request of the sponsor.
SECRET
12
SECRET
F-j
Declassified in Part - Sanitized Copy Approved for Release 2012/08/30: CIA-RDP79B00873AO01600040034-7
 Declassified in Part - Sanitized Copy Approved for Release 2012/08/30: CIA-RDP79B00873AO01600040034-7
SECRET
8
a
a
i
i
i
B
i
A
B
0
B
8
SECRET
Declassified in Part - Sanitized Copy Approved for Release 2012/08/30: CIA-RDP79B00873AO01600040034-7
a