PROJECT RINGAROUND

• Document Type: CREST
• Collection: NGA Records (Formerly NIMA)
• Document Number (FOIA) /ESDN (CREST): CIA-RDP78B04767A000100160001-8
• Release Decision: RIPPUB
• Original Classification: T
• Document Page Count: 17
• Document Creation Date: December 28, 2016
• Document Release Date: June 10, 2003
• Sequence Number: 1
• Publication Date: November 16, 1967
• Content Type: MFR

Approved F
r Release 2003/07/14: CIA-RDP78B04767A000100160001-8
16 November 1967
25X1
MEMORANDUM! FOR THE RECORD
SUBJECT: Project Ring Around
REFERENCE: TA-37-500019-2, 3 October 1967
NRO REVIEW COMPLETED
copy---.L--
1. Color film has been used in various collection systems,
in both operational and test missions. Reproduction copies can
vary widely in color balance, depending upon the processing and
desired results for interpretation. To formulate a meaningful
PI opinion on the color gradation and density of duplicates
preferred which may be used as standards for color duplication,
NRO initiated Project Ring Around. The processing facility
provided a series of color reproductions to NPIC. In NPIC,
both PAG and TID conducted a coordinated test to determine the
preferred reproductions for exploitation.
2. The referenced document is a report of the test, pre-
pared by the processing facility. The test produced beneficial
data that will govern reproduction standards in the future. As
a result of the test, it was concluded that:
0
ductions.
Attachment : A/S
b. Less density is desired.
c. A slight shift in the "aim point" color balance
is preferred.
25X1
25X1A
25X1A
Declass Review by
NIMA / DoD
is an acceptable emulsion for color repro-
Depu. y ie
Technical Intelligence Division
Distribution:
Copy 1 - NPIC/O/Dir
2 - NPIC/Asst for PA
Copy 3 - NPIC/Asst for TD
4&5 - NPIC/TID
25X1
'Approved For Release 2003/07/14 :. CIA-RDP78B04767A00010016
 Approved For Release 2
0001-8 NRO 25X1
25X1A
This document contains TA-37-500019-Z
13 pages
Copy of 30
DATE: 3 October 1967'
25X1A
NRO 25X1
Approved For Release 2003/07/TPPCI4 8B04767A000100160001-8
 Approved For Release 2
UaIU 1/ 114 :-CIA-KUV f 8b 717
1 - 7
20 SEPTEMBER 1967
TA 37-500019-a -
NRO 25X1
20 September 1967
NRO 25X1
Approved For Release 2003/07/T4JtPC1B04767A000100160001-8
 Approved For Rele
e 0100160001-8
TOP SECRET
TA 37-500019-
SPECIAL REPORT 20 SEPTEMBER 1967
TITLE: Project Ring-Around
TASK/PROBLEM:
NRO 25X1
1. Provide for nhotointerpreter evaluation, a series of aerial
color photographic reproductions showing graduated color and density varia-
tions, from which preferences may be subjectively selected. Analyze these
data and translate into objective terms which may be used as standards for
color aerial reconnaissance duplication tasks.
2. Reproduction standards for color aerial reconnaissance
photography have heretofore not been specified with universal agreement.
Color balance of the duplicate transparencies has been particularly trouble-
some, with lesser disagreement on highlight or overall density. In an effort
to determine what the ultimate user of the material requires, or at least
desires, series of duplicates were prepared which contained controlled color
and density variations. Ten typical aerial scenes, selected by the customer,
were used to prepare each "color ring-around" and density series.
3. Using a standardized test procedure, 3L photointerpreters
evaluated the coded test duplicates and se tested individual preferences for
color and for density. These subjective data were then reduced to objective
terms by means of small-spot integral densitometry.
4. Test scenes selected by the customer were largely industrial-
residential complexes. The preferred overall density for the average of
these scenes, and the average of all viewers, was 0.90 D as measured through
the red filter of a standard Type A transmission densitometer. The preferred
color balance at this red density is 1.18 D through the green filter and 1.55 D
through the blue filter. Because the measured color balance changes for sub-
jects at other red-filter densities, a calibration curve has been prepared
relating the preferred three colors.
5. The appropriate color balance for reproductions of aerial
reconnaissance photography has long been a subject for conjecture. Because
the original photograph is often, of necessity perhaps, of a color balance
quite far from "normal" in the personal-photography sense, some have believed
that the reproduction color balance need only lie anywhere between that of
the original and "normal". For mensuration purposes this may he true. How-
ever, there is increasing evidence that photointerpreters in general need, or
at least desire, a color balance which will show all commonly-recognizable
objects in w hat is considered to be their "true" color. It is the purpose
NRO
Page 2 of 13 25X1
Approved For Release 2003TOPI4 SECU DP78BO4767A000100160001-8
 -1 R R -3 F -P F__ 2t
-.00 Approved For Releas - 0100160001-8 25X01
TA-37-500019- Z
TOP SECRET
20 SEPTEMBER 1967
25X1 D
of the study herein reported to survey the customer wishes on both color
balance and density of aerial photography reproductions, and to translate
such subjective information into objective terms of immediate practical
usability. Other independent studies of a related nature now underway
should, in the future, lead to more sophisticated equipment and procedures
to accomplish the same ends on an automated mass-volume basis. For the
present, standard manual three-color densitometry is used and can provide a
large measure of standardization of this troublesome subject.
6. High altitude aerial photography is often hampered by the
presence of haze, smog, dust, etc. which has the effect of filtering out
and scattering some of the light which should make up the total imagery.
Filters are commonly used to penetrate these interfering foreign layers,
but they introduce their own distortion of the original imagery. The ultra-
violet absorbing filter usually used for color photography results in a
yellow color balance on the original film which is very unnatural in terms
of typical amateur photography. However, if sufficient red, green, and blue
light has reached the film during exposure, it may be theoretically possible
to selectively print for color balance the original film so that the three-
color relationship for readily-recognizable subjects is brought back almost
to that which is normally perceived at ground-level viewing. Practical con-
siderations limit the degree to which this may be accomplished, and there
has been considerable disagreement on the need or desirability of doing so
in the first place.
7. Tone reproduction studies and other purely objective proce-
dures should ultimately spell out the answers to the current questions on
acquisition and reproduction requirements. Until some of the inconsistencies
of some of these evaluation techniques can be resolved, and the interrelation-
ships better understood, the subjective requirements of the ultimate users of
the aerial photography must fill the void. We have attempted in this study
to provide a large group of these ultimate users a selection of typical color
reconnaissance photography from which individual preferences for color balance
and density could be subjectively selected. The results of their efforts
have been compiled and translated into the one objective procedure which is
currently readily available: three-color manual integral densitometry. More
sophisticated methods and instrumentation are in the offing, but are not yet
suitable for practical application.
I I8. A typical high-altitude color original film, identified as
was selected by the customer for use in this study. Imagery on
is i. m was judged to be of high quality, but because of the camera filtra-
tion which had been applied, the color balance was roughly equivalent to at
least a CC .L10 yellow to yellow-green from ground-normal. The duplicate which
NRO
Page 3 of 135X1
Approved For Release 2003NOA4 SMEDP78BO4767A000100160001-8
 25X1 D
25X1 D
25X1 D
Approved For Relea
00100160001-8
TA 37-500019-.9-
20 SEPTEMBER 1967
originally had been supplied to the customer had some compensation for this
excessive yellow balance, but was still judged to be approximately CC .20
yellow from normal. It has not yet been established the extent to which
this condition restricted or otherwise hindered the photointerpretability.
9. Using original film l a new color balance for dupli-
cation was determined which would appear more nearly "normal". This was done
first by purely objective means through densitometry of concrete objects for
which considerable reflectance and transmission density d-ta had previously
been generated. Because this color balance appeared bluer than was considered
to be "normal", a shift toward a yellower color balance was subjectively
induced. This new "reference" color balance is approximately CC .20 bluer
than the duplicate originally supplied to the customer. Using the reference
as the mid-point of a color series, duplicates were made which deviated by a
CC .05, .10, and .20 in each of the six color directions; i.e., red, green,
blue, cyan, magenta, and yellow. Ten scenes as selected by the customer were
so prepared and sent to him for evaluation. Additional density series, at
a single color balance, were similarly prepared.
NRO
25X1
10. The test scenes selected by the customer contained a wide
variety of subject matter. Arid desert land, temperate forest and agricultural
areas, and semi-tropical and coastal lands were represented. Included in these
formats were cities and towns, industrial and rail facilities, and other spec-
ific targets of the type which might be of strategic interest. Due to the
acquisition system used, each est sample represented upwards
of Iland suriace.
11. The basic scheme for evaluating the color variation test mate-
rial was to have each of may photointerpreters view a single sample under
7X magnification, and to similarly intercompaare several samples. By the pro-
cess of elimination and preferential selection, each photointerpreter would
determine his first, second, etc., choices for color balance and for density,
for each of the test scenes.
12. When the project was presented at the customer's shop, two
practical problems soon became apparent. First, the overall color -' a large
format can subconsciously influence a viewer's judgment of detail color, so
that unimportant imagery can induce false evaluation of target imagery.
Second, it is difficult to shift one's viewing under magnification from one
sample to another, and retain adequate detailed judgment so that consistency
can be maintained. The farther removed two test samples are, physically, the
more difficult is a valid evaluation.
13. Because of the practical difficulties encountered, it was
decided by the customer that the test samples be reduced in size. Accordingly,
out of the original ten large formats, thirteen specific targets were selected.
These were extracted as one-inch square film chips, which were numerically
coded as had been the original large formats. The chips were mounter] in ran-
dom sequence on transparent sheets in blocks of three chips across a row and
three rows per grouping. It was thus easy to intercompare the image charac-
NRO
Page 4 of 15X1
Approved For Release 2003/Q : SMAR$1P78B04767A000100160001-8
 Approved For Rele se - A000100160001-8
TA 37-500019-Z
TOP SECRET
20 SEPTEMBER 1967
NRO
25X1
teristics of various color balances because of the proximity of the selected
samples.
14. The detailed procedure by which the entire test was conducted
is contained in written instructions provided to the customer (Appendix I).
Some of the nar.ameters which received particular attention were: predetermina-
tion of each photointerpreters color perception, standardized viewing condi-
tions (same magnification, illumination source, inspection sequence, etc.,)
and coded sample identification to avoid color-label influence. Therefore;
within practical limitations, this was a "controlled" test.
15. There was one deviation from the plane=d test procedure which
the customer elected to retain: the transparency illuminators in the customer's
shop have a color temperature averaging around 103000?K instead of the recom-
mended 14,1400?K. Because of the fluorescent light source, which does not have
a continuous spectral emission, the indicated color temperature is only an
approximation. The apparent color is considerably bluer than the recommended
source. This discrepancy is acceptable as long as the illumination remains
as defined and indiscriminate intercomparisons are not attempted.
16. Thirty-four photointerpreters in the customer's shop performed
the prescribed evaluation of the thirteen series of color balance and density
variations. Each one indicated his first through fifth choices for color
balance, and first through third choices for density, for each subject. The
preferences were expressed as "chip numbers", because no other identification
was provided during; the test. These data were then summarized into frequency
tabulations, which were forwarded to the contractor for evaluation. (Fig. 1)
17. The entire array of one-inch film chips was also returned to
the contractor. Because it was the goal to relate customer subjective prefer-
ences to some objective evaluation, the most frequently selected chips were
submitted for densitometric determinations. Manual integral densitometry is
the only system currently available for immediate practical lication.
Therefore, the selected chips were read on densitometers, 25X1
using Status "A" filters and both lmm and 14mm apertures. ep ication was
employed to improve the validity of the data so generated. A specific small
spot was preselected from within each one-inch square subject area, and every
effort was made, manually, to measure this same spot each time densitometry
was performed. Because each millimeter of film displacement in the densitom- 2 5X1 D
eter represents approximately f ground subject matter, any slight
error in repositioning the chip could result in a significant change in the
imagery being measured. As in actual mission experience, however, there is
no assurance that any single spot is more representative of the scene than
any of the adjoining spots. This inherent variability, therefore, is trouble-
some but an inescapable part of aerial photography evaluation.
I. The customer-selection frequency tabulations (Figure 1) show
that no single color balance received a clear majority of preference. Because NRO
25X1
Approved For Release 2003/9OP4 : QRTP78BO4767A000100160001-8page 5 of 13
 Approved For Rele
TA 37-500019- Z
20 SEPTEMBER 1967
deviations of .05 Log E are difficult to perceive under magnification, some
of this variability was to be expected. Those preferences which deviated
from the apparent mean by more than this amount, and particularly in a color
direction contrary to the general trend, further emphasize the imperfect
human element in this type of study.
NRO
25X1
19. When the customer-selection data are decoded into Log E
deviations from the reference, the first, second, etc., choices can be graph-
ically shown as on Figure 2. In spite of the variability mentioned above,.
it is interesting to note that over two-thirds of the first three preferences
fall within .05 Log E of the reference, and that there is a strong indication
of preferred color direction. The data, therefore, do assume meaningful pro-
portions. Note that each major division on the graph (Fig. 2) is only 0.05 Log E.
20. The densitometric data can also be presented in a graphical
manner. Different subjects, even of the same general type such as "factories",
will have different total densities. Furthermore, the relationship of red
to green to blue light reflectance changes for various overall densities.
Therefore, for each measured chip of a given preference level ("first choice",
for example), the red-filter density was used as the abscissa for the graph-
ical presentation, and the magnitudes of the blue-filter and the green-filter
densities above the red filter density were the ordinates. All of the points
so obtained for the Reference Chips are shown for illustrative purposes in
Figure 3. The smooth curves drawn through the scattering of points represent
the best estimates of the "true" relationship of the three colors.
21. In a manner as described above:, the first three customer
choices for color were individually plotted, treating the lmm and the 4mm
densitometer aperture data separately. These six graphical presentations were
so similar that a single composite graph has been drawn, as shown in Figure 14.
This, then, is the calibration which describes the average customer color
preference for the types of subject matter used in this study, in terms of
the mode of integral densitometry currently used in the contractor's shop.
22. Densitometry of the density-preference film chips did not lead
to as well-defined a correlation as did the color study. Statistical evalua-
tion of the data was made, and a set of densiometry values was extracted as
shown on the color-calibration curves, Figure 4. However, because the subject
matter available for this study varied so much in "minimum" and in "average"
density, the figures presented here as tentative aim values must be viewed
as "first approximations" only. Subjective adjustment may be necessary in
actual mission application.
23. Thirty-four experienced photointerpreters have evaluated color
and density variations introduced into the reproduction of 13 aerial color
reconnaissance scenes. From their subjective conclusions, objective data have
NRO
Page 6 of 13 25X1
Approved For Release 2003/VCp4 :SIBVP78B04767A000100160001-8
 Approved For Re
-
TOP SECRET
been extracted which specify the preferred color balance and density for
similar photography. When expressed as Status "A" integral density values,
these data provide the standards, within approximately .05 Log E, by which
future aerial color reproductions may be made. These data also may be
directly applicable to, or at least serve as approximation for, other acqui-
sition systems where scale factors, subject matter, acquisition mode, etc.,
are quite different.
24. Duplication specifications, as determined by this study, may
be restated as follows:
Original Film:
Duplication Film
Process: Modified(lowered contrast for
00100160001-8 NRO
TA 37-500019-2- 25X1
20 SEPTEMBER 1967
Densitometry: Status A filters, lmm or !rnm aper-
tures
Aerial Subject matter to be measured: Industrial or residential com-
plexes
Average Preferred Densities:
Red filter 0.90
Green filter 1.18
Blue filter 1.55
25. For subjects lighter or darker than the "average", refer to
Figure 4 for the curves relating the three color preferences.
26. Verification of the conclusions of this study should be made
by utilization wherever color aerial film duplications are to be made. By
documenting failures as well as successes, refinement in the technique can
be made. Similarly, effects of scale factors, available subject matter, etc.,
can be noted for further validation of the procedure. Assuming that funda-
mentally the approach used here is valid, it will be essential to pursue
advanced instrumentation if mass-production will ultimately be required.
25X1A
25X1A
25X1
Page 7 of 13
NRO
25X1
Approved For Release 2003/Q0 : $Uf7P78B04767A000100160001-8
 Approved For Relea
TOP SECRET
APPENDIX I
A000100160001-8
TA 37-500019- 2~
20 SEPTEMBER 1967
Suggested Instructions for PROJECT RING-AROUND Preference Survey:
Basic Provisions:
1. Provide a viewer or viewers capable of handling at least five
transparencies.
2. The color temperature of the viewer should be 4400?K ? 200?K and the
intensity of the light source should be 450 ? 50 foot candles.
3. It is recommended that masks be used to reduce extraneous light from the
viewer.
4. Incident room light should not exceed 10% of transmitted light from
viewer.
5. Evaluation should be done at 7X magnification.
6. In order to obtain meaningful data from this color preference survey,
a simple screening test should be given to separate persons with defec-
tive color vision from those with normal color vision.
NRO
25X1
25X1A
It is suggested each participant in the survey be screened for color
blindness by customer personnel using the 25X1
Pseudoisochromatic Plates Test and procedures.
If a color blind person is included in the survey, it should be so noted
on the score sheets so that his evaluation can be given appropriate weigh-
ing in the statistical analysis of the data collected.
7. A minimum of 10 evaluators with normal color vision are suggested to
evaluate all the samples in the ring-around.
8. All. evaluators should be given identical instructions.
9. Each evaluator should independently judge the series.
Procedures
1. Date the score sheet. Assign a number to each evaluator and record in
upper left corner of the test sheet, also indicate color blindness -
Y he is color blind or N he is not color blind.
2. Select one of the ten boxes of duplicates.
3. Note revolution and frame number and remove all the duplicates in the
30 series.
NRO
Page 8 of 13 25X1
Approved For Release 20031110P4 : 514J7P78B04767A000100160001-8
 Approved For Release
TOP SECRET
APPENDIX I (continued)
000100160001-8
TA 37-500019- -Z
20 SEPTEMBER 1967
Instruct the evaluator to examine under 7X magnification all 21 color
duplicates and select' 10 most acceptable for color balance. Then rank
these 10 in order of his color preference.
Evaluators may use different methods of ranking for color preferences.
The method used is immaterial, providing a lst, 2nd, 3rd, etc. prefer-
ence is indicate d.
5. Record under the appropriate column of the score sheet, in order of his
preference, the number following the words PROJECT RING-AROUND.
6. Ask the evaluator to indicate which, if any, of the 21 duplicates are
acceptable for color balance.
7. Circle those numbers so indicated on the score sheet.
8. Remove the 21 duplicates and display the 7 duplicates in the 30 series.
9. Ask the evaluator to judge density in order of preference and record
these 7, two-digit numbers on the score sheet in the density column.
10. Follow procedures 2 thru 9 for each of the other sets of duplicates.
11. It is suggested the same evaluators repeal; the exercise at a later date,
a day or so later, using the same procedure as outlined above.
NRO
25X1
Page 9 of 13 NRO
25X1
Approved For Release 200311OP4 : l@RIRpP78B04767A000100160001-8
 Approved For Releaso
2003/07/14: CIA-RDP78B04767A 00100160001-8
EXAMPLEr
FREQUENCY OF CUSTOMER
SELECTION: 1st, 2nd, .3rd,
4th, and 5th CHOICES
01
z~,s'~~L~ 11)s
`~ ~ 31y33z;
02.
.,03
04
05
333y 'fIS7 9 --
.06
IL I
07
08
TA 37-500019- 0.
R
1
.05R
.05G
.20C
3
z
.o5c
7
5
.10Y
.20'
FILM CHIP IT
FICATION COT
ENTI..
E
.1OR
.10B
y
'
I
2?
.20Y
3
!
1
3
3
3
2
Z I
/
3
!
~
?05B '
N
T
.
% t
.
.20M
r
~
~
s
3
i
I
V
3
y
.20R
)I
3
Z
'~
~
I
' '
J
REF
1
I
3'
1
~
_
~
.05M
Oy,
~
3
J
)
J
,.
~
CO
DE
LO
VI
R
AT
BA
IO
LA
N
NC
FR
E
OM
.100
11,>
4
,
y
z!
~
RE
FE
2E
NC
E
.05Y
s:. Z
IL
L
Z
Z Z.
7-
2
3
..I,
+.10
7- 2-
Z
LJ
.
+
05
-
-
-
1
I I
Approved For Release t 7A000100160001-8
NRO
25X1
NRO
25X1
 Approved For Release 2003/07/14: CIA-RDP78BO4767A000100160001-8
25X1A
BALANCE OF OREPROIACTIONS OF
AERIAL RECONNAISSANCE PHOTOGRAPHY
NRO
25X1
CUSTOMER PREFERENCES:
. First Choice
o Second Choice
x Third Choice
NOTE: THE "REFERENCE" COLOR
AS THE MID-POINT FOR THE COLOR
4-12
OF X
0.05 Log E
Approved For Release 2003/07/14: CIA-RDP78B04767A000100160001-8
 Approved For Relea
e 2003/07/14: CIA-RDP781304 67A000100160001-8
TA 37-500019-i
~''Mmrmrnmmn~mm'"
-------------------------
EffliLi
IN
Id I
t
URM 11111
ITflufli", jjjj~j [loll,
liul
11I~IiffNIM[HfflHfflllflhIflUffl
~rinrrmni~mmmmmrl n~oir, u r uc
~rmu~u~u!~mr~n~~rm~mr
DENSITOMETRIC EVALUATION OF
COLOR "REFERENCE" CHIPS FROM
IE:) .
1ENSI-
TOME'STA US "All FILMS
NRO
25X1
25X1
NRO
25X1
Approved For Release S E 67A000100160001-8
 Approved For Releas
TA 37-500019- a-
COLOR BALANCE AND TENSITY PREFERENCES FOR
REPRODUCTIONS ON I FILM OF COLOR AERIAL
RECONNAISSANCE PHOTOGRAPHIC ORIGINALS
IIENSI TOMETE R
STATUS 'tn- r ILTERS
Page 13 of 13
Approved For Release 000100160001-8
NRO
25X1
25X1A
25X1
NRO
25X1
 CENTER ROUTING SL2 2 L3!_
FROM Approved or a ease
C177t;-
ATE
/6 Nou 67
TO
INITIALS
DATE
REMARKS
DIR
r
wk
10- 4
Scr,t o got, 4 r -'p
rmaCIok
DEP/DIR
4
!e `~
an ~~. I. (
c~ 3 Ott I~Y+~cu
EXEC/DIR
3
(f
p-o a A- h_j Arota.KcL has
TECH ADV
2
/, (
ja~Ch C.ov+ti~ z..d-
SPECIAL ASST
~t?cer-
CH/S5
ASST FOR P&M
ASST FOR OPS
ASST FOR PA
t l~ 7z-C G''
ASST FOR TD
I
CH/CSD
CH/IPD
CH /PD
y' ;`i9w L2~/f
CH/PSD
CH/TID
CH/IAS/DDI
C H %D IA /XX -d
CH/DIA/AP-IP
"-
CH/SPAD
LO/CGS/CIA
LO/NSA App
rove
d FUF
Vleas
2003/07/14: CIA-RDP78B04767A0001
25X1A
I P FM 30 (2-67) DESTROY PREVIOUS
EDITIONS
 25X1 Approved For Release 2003/07/14: CIA-RDP78BO4767A000100160001-8
Approved For Release 2003/07/14: CIA-RDP78BO4767A000100160001-8