PROJECT 6619

• Document Type: CREST
• Collection: NGA Records (Formerly NIMA)
• Document Number (FOIA) /ESDN (CREST): CIA-RDP78B04747A001100030043-0
• Release Decision: RIPPUB
• Original Classification: K
• Document Page Count: 5
• Document Creation Date: December 28, 2016
• Document Release Date: November 26, 2001
• Sequence Number: 43
• Publication Date: May 9, 1969
• Content Type: LETTER

Approved F ,Release 2002/05/08: CIA-RDP78B04WA001100030043-0
May 9, 1969
Washington, D.C.
Subject: Project 6619
Dear John:
Enclosed are two (2) copies of Progress Report
No. 6 in accordance with the schedule of the subject
contract.
Sincerely,
Program Manager
WWM/mis
encl : As stated
Declass Review by NIMA/DOD
STATINTL
Approved For Release 2002/05/08 : CIA-RDP78B04747A001100030043-0
 Approved Frelease 2002/05/08 : CIA-RDP78B04Z 7A001100030043-0
PROGRESS REPORT NO. 6
COLOR IMAGE ASSESSMENT
PROJECT 6619
by
Period: April 1 through April 30, 1969
Approved For Release 2002/05/08 : CIA-RDP78B04747A001100030043-0
 Approved For Release 2002/05/08: CIA-RDP78B04gA001100030043-0
PROGRESS DURING THE PERIOD
The modulation transfer function phase of this project was
initiated with delivery of three analytical color friskets on
S.O.151 camera original emulsion. The emulsion had been "exposed
out" to yield single analytical dye layers. Thus, of the three
friskets, one was exposed only in the yellow dye layer, one in the
magenta dye layer, and the third in the cyan dye layer. A descrip-
tion of the frisket was provided in the previous progress report.
The trace instructions were as follows:
1. Resolution readings should be visually taken from all
resolution targets.
2. A scan across the micro-step wedge with a 1 x 80pu slit
should be performed with a 10, sample interval.
3. All edges in the first row (all orientations and all
contrasts) should be scanned with a 1 x 80,,a slit, using
'a sample interval of 1 .
4. All edges in the fifth row (all orientations and all
contrasts) should be scanned with a 1 x 801.L slit using a
1 y. sample interval.
Once these traces are received from the customer, utilization
of the exposure table generation routines and the Fourier transform pro-
gram will yield the analytical modulation transfer function.
The direction cosine routine has been utilized to generate polar
plots of the S.O. 121 dye system and the S.O. 151 dye system. The
results indicate that Beer's law holds for all combinations of the
dye layers for each emulsion. The angle corresponding to the direc-
tion cosine computed in the program indicates the degree of inde-
pendence with which one dye layer in a tripack absorbs relative to
Approved For Release 2002/05/08 : CIA-RDP78B04747A001100030043-0
 Approved F,Release 2002/05/08 : CIA-RDP78B047.a7A001100030043-0
either of the other two layers. Completely independent dye systems
yield an angle of 90? and are termed orthogonal.
The following table lists by film the basic results of the
direction cosine analysis:
Layers
Angle
YC
78.2?
YM
70.2?
CM
65.0?
Layers
Angle
YC
76.0 0
YM
69.5?
CM
66.7?
Emulsion number 121-41-31
Layers
Angle
YC
80.0?
YM
70.0?
CM
64.0?
The exposure generation routine has been completed and tested.
Errors in the regenerated exposure table have appeared. Analysis
of the output in order to determine the origin and nature of these
errors is proceeding. One.problem involves the generation of the
variance - covariance matrix directly from a deviation matrix. This
may be an incorrect procedure mathematically. A routine is presently
being tested that removes both the neutral and the average deviation
of the neutral from the set of analytical exposure tables. Also, a
modification to the sets of exposure tables used in the VECTOR pro-
gram appears to reduce exposure table generation errors. This
Approved For Release 2002/05/08 : CIA-RDP78B04747A001100030043-0
 Approved F4P
,0elease 2002/05/08: CIA-RDP78B04747A001100030043-0
modification consists of fixing the value of the cut off point on
the toe and shoulder of each exposure table to a constant exposure.
This effectively reduces the dimensionality of the variance - covar-
iance matrix in the VECTOR program.
Testing of the regeneration capabilities of a set of character-
istic vectors for exposure tables generated from colors outside the
original set of system color (colors in the original set of
exposure tables used to generate the set of characteristic vectors)
indicates that no problem exists. Non-system exposure tables may be
regenerated with accuracy equal to system exposure tables providing
the color domain of the original set of system exposure encompasses
the domain of the set of nonsystem colors (i.e., the set of original
system tables wholly encloses the set of non-system tables).
Work during the upcoming month will include final testing of
and criteria selection for the color exposure generation routine
(SELGEN), running of the MTF routines providing satisfactory traces
are received from the customerp and initial drafting of the final
report for the contract. A trip to the customer's facility will be
made to discuss possible problem areas in the implementation of the
programs.
Approved Fbr Release 2002/05/08 : CIA-RDP78B04747A001100030043-0