ANALOG IMAGE RESTORATION

• Document Type: CREST
• Collection: NGA Records (Formerly NIMA)
• Document Number (FOIA) /ESDN (CREST): CIA-RDP79B00873A000100010186-8
• Release Decision: RIPPUB
• Original Classification: S
• Document Page Count: 8
• Document Creation Date: December 28, 2016
• Document Release Date: November 1, 2012
• Sequence Number: 186
• Publication Date: June 24, 1966
• Content Type: MISC

Declassified in Part - Sanitized Copy Approved for Release 2012/11/01: CIA-RDP79B00873A000100010186-8
RIP
ANALOG IMAGE RESTORATION
1. PROBLEM:
To extract more information from degraded imagery.
2. FACTS BEARING ON THE PROBLEM:
a. All photographic imagery is degraded to some extent by
acquisition system limitations and by atmosphere-induced aberrations.
b. Even the smallest aberration in a critical portion of the
imagery can be immeasurably,expensive in terms of information loss.
c. This loss will be increased by the higher resolution imagery
from new c4lection systems.
d. Efforts to improve the quality of imagery after acquisition
are hampered by a lack of basic scientific information.
2 4 JUN 1966
e. under contract for EPIC, has
recently aemonstrated the feasibility of image restoration through
the technique know as spatial filtering.
f. This project for FY 1967
phase of a successfully completed
project, the :technique of spatial
and experimentally evaluated.
DISCUSSION
a. Origin of Concept.
This follow-on project is part of the Image Analysis Category)
one of the 1 major categories forming the NPIC Technical
Development Program. The project was initiated in DPOPMIlemy. 10),
by an unsolicited proposal from 25X1
to study the feasibility of restoring imagery by optical processing
techniques. On the basis of that proposal, a one-year contract ncN/4
was negotiated with
9)(1
The contract cost
and was in effect from 15 May 1965 to25X1
15 May 1966. ,
is a planned specific follow-on
FY 1965 contract. Under this new
filtering will be further developed
25X1
Under the contract,
use of coherent light"DroppgRin
introduced
the construction or holograms for use
masks, to be placed over the degraded
performed to date has shown that this
is possible to express mathematically
-En
investigated the
r.inipensate for the distortions
This technique involves
as spatial filters, or
imagery being viewed. Work
technique is feasible. .It
the aberrations incurred in
-SteR-E-1_
1 ----1,11U1
Exclude
gYll'?:?Vi53; Anti
ii8daISHICI:.11:1 ,
Declassified in Part - Sanitized Copy Approved for Release 2012/11/01: CIA-RDP79B00873A000100010186-8
f
25X1'
25X1 ?
c'Cf.17(
Declassified in Part - Sanitized Copy Approved for Release 2012/11/01: CIA-RDP79B00873A000100010186-8
a simg/photographic system and in anikerimental environment
it was possible-to 'construct a spatial filter which corrected
certain aberrations. Theoretical studies indicated that other
distortions can be corrected in a similar manner.'
b. Proposed Program.
The studies and experiments to date have each involved a
single type of aberration. The results have indicated that
the same techniques can be applied to the complex aberrations
foUnd on operational imagery. The next phase, to be carried
out under the proposed contract, is intended to validate this
theory by the experimental construction and testing of spatial
filters.
c. Selection of Contractor.
During the past year, , has
exhibited an outstanding ability to understand NPIC problems
and has made tremendous progress in an area not previously
explored by the intelligence community.
submitted an unsolicited proposal to continue the
on the basis of its progress to date and the high
the new proposal, the company was chosen for this
d. Program Phasing.
study and,
quality of
contract.
As shown graphically on the attached TAB C, there will be
major milestones at the end of two, Six, and twelve months in
this one-year project. At each of those' points, NPIC technical
monitors and Agency decision makers will be able to decide '
whether the goals of the project are lbeing met- and whether
further efforts should be abandoned, redirected, or continued- -
as planned. ' 'An estimate of the three-year 'effort necessary
to make this technique operational is illustrated in TAB D,
e. Coordination.
25X1
25X1
There has been thorough coordination of this ?ro ect with
DD&ST through discussions with of ORD. In 25X1
addition, there is similar coordination with t e Air Force the
only other government component known to be sp nsoring research
in this field.
f. Alternatives.
This project is similar in ultimate purpose to the Digital
Image Restoration Project we are supporting at 25X1
However, there are fundamental differences in the basic approach
and the potential applications in each case. The 25X1
project appears to be much closer to the practical application phase.
4. CONCLUSIONS:
The Image Restoration project holds promise of a major breakthrough
capable of vastly enhancing the photo interpretation process by making
S::CPT"7"I
Declassified in Part - Sanitized Copy Approved for Release 2012/11/01: CIA-RDP79B00873A000100010186-8
V
Declassified in Part- Sanitized Copy Approved forRelease2012/11/01 : CIA-RDP79B00873A000100010186-8
111 .
SECRET
'available information which can be obtained through no other means.
ResUlts to date have been highly satisfactory and the contractor
'involved has demonstrated 'outstanding competence.
5. RECOMMENDATIONS:
IL is recommended?that abloroVal bp granted for .a
with at a funding level of
for 'Iscal Year 1967.
f011bw-on contract
6. REFERENCES AND ATTACHMENTS:
TAB A.
TAB B.
TAB C.
TAB D.
Catalog Form
Technical Discussion
First Year Program Phasing
Three Year Program Phasing
SECRET
3
P
025X1
L....1/? I
Declassified in Part - Sanitized Copy Approved for Release 2012/11/01: CIA-RDP79B00873A000100010186-8
Declassified in Part - Sanitized Copy Approved for Release 2012/11/01: CIA-RDP79B00873A000100010186-8
When F
3 INF
R & D CATALOG FORM
DA TE
22 June 1966
I. PROJECT TITLE/CODE NAME
Analog Image Restoration
(Spatial Filtering)
2. SHORT PROJECT DESCRIPTION (Follow On)
Correction of the phase of a transfer function
of an aberrated image system.
3. CONTRACTOR NAME
4. LOCATION OF CONTRACTOR
25X1
5. CLASS OF CONTRACTOR
6. TYPE
Research Laboratory (Commercial)
OF ,VIIIMAloi
CPFF
7. FUNDS
13. REQUISITION NO.
9. BUDGET PROJECT NO.
FY 19 66 $
25X1
NP-A-3-8315
FY 19 67 $
10. EFFECTIVE CONTRACT DATE
(Begin - end)
11. SECURITY CLASS.
AA-Secret
FY 19 8 $
T-Unclassified
W-Secret
12. RESPONSIBLE DI
L
PROJECT OFFICER TELEPHONE EXTENSION
DDI/NPIC/P&DS,
25X1
, ?
13. REQUIREMENT/AUTHORITY
' The continuation of this study is essential to int-reage
in deriving valuable intelligence information from degraded
NPIC's effectiveness
imagery.
14. TYPE OF WORK TO BE DONE
Investigation in the problem areas of optical data processing
to compensate for degradations introduced by the low-pass-frequency
an optical system.
techniques
nature of
IS. CATEGORIES OF EFFORT
MAJOR CATEGORY
SUBCATEGORIES
.
Optical Systems
Image Analysis
Photographic Enhancement
Image Restoration
16. END ITEM OR SERVICES FROM THIS CONTRACT/IMPROVEMENT OVER CURRENT. SYSTEM. EQUIPMENT.
Reports on methods and techniques for constructing a filter
for amplitude and phase distortions.
_
ETC.
to correct
?
17. SUPPORTING OR RELATED CONTRACTS (Agency & Other)/COORDINATION
I
Thorough coordination with DDS&T has been accomplished
representatives. Coordination with the Air Force to prevent
through ORD
duplication of
the following contracts has also been accomplished:
25X1
25X1
IS. DESCRIPTION OF INTELLIGENCE REQUIREMENT AND DETAILED TECHNICAL DESCRIPTION
tional page if required)
The purpose of this program is to apply coherent opti
techniques to the restoration of photographic images. Dur
program the transfer function of an optical system was det
graphic techniques, the inverse of the function was constr
distortions in a system have been corrected using the invelse
work will be directed toward correcting phase distortions using
OF PROJECT (Continue-on addi-
al processing
ng first year of this
rmined;then by photo-
cted. The amplitude
filter. Future
similar techniques.
19. APPROVED BY AND DATE
/
OFFICE
1
1
--__
DEPUTY DIRECTOR
I:MCI
0104 ert113C7 I ma.,
Declassified in Part - Sanitized Copy Approved for Release 2012/11/01: CIA-RDP79B00873A000100010186-8
Declassified in Part - Sanitized Copy Approved for Release 2012/11/01: CIA-RDP79B00873A000100010186-8
TAB B ?
, TECHNICAL DISCUSSION
A brief discussion of the procedure using a hologram as a means of
preserving phase information in the image restoration process follows:
Since inverse filtering is a linear process, a linear systems analysis
will be assumed here. It is known that an incoherent image abberated by :a.
linear system may be represented as:
where
Iim(y) =ob(x) s(y - x) dx
Iob(x) = the object intensity distribution
s(x) = the impulse response of the system.
If the impulse response of the system is symmetric, then its Fourier
transform (i.e. the transfer function) is real. This case has been 'Studied
and in some instances corrected images have been obtained by constructing
inverse filters of real transfer functions. The problem of interest here
considers systems in.which the impulse response is not symmetric. However,
it is important that the impulse be stationary (the shape of the impulse
response not change as the point source in the object plane is moved about).
This requirement is necessary because we want the same filter to correct
the entire photographic frame.
First a hologram of the system transfer function, which in general
will be complex, is made. This is done by optically Fourier transforming
the transparency containing the impulse response with a coherent beam.
The hologram will be made by mixing the transfer function with a plane wave
at an angle,oco, to the axis. This may be done by utilizing Fourier transform
holograms as shown in Figure 1. The intensity distribution in the hologram
is given by
ika x
1(x) = le ? + r(x)1 2
ornnrr
Declassified in Part - Sanitized Copy Approved for Release 2012/11/01: CIA-RDP79B00873A000100010186-8
i,
-----N.Ti--"4
Excluded fro auloail.
donzradin and
.....".
(2)
Declassified in Part- Sanitized Copy Approved forRelease2012/11/01 : CIA-RDP79B00873A000100010186-8
?
whereT(x) denotes the transfer function of interest. Next a transparency'
varying as 11/r(x)12 is prepared.
ie
Beam k f f
Splitter . Objeotg) Hologram (X)
Figure 1_ Fourier Transform Holograms
If Eq. (1) is prepared as an optical transparency and then Fourier
transformed optically, the aerial image amplitude is described by
9craW = riob(x) r(x)
If the hologram transparency of Eq. (2) and the 1/17(x)12 transpar&Icy are
sandwiched together and placed in the aerial image of Eq. (3), the result is
(3)
ob(x) r(x) ika x -ikaox)
(x) ?I 2 *
+ I Ttx/ e o +re
im
r*(x) r(x)
The third item in Eq. (4) gives
A.3 ika x
im(x) = Tob(x) e
which upon Fourier transformation by a lens of focal length f gives
a
I 01im(Y)
ob Xf X
(4)
(5)
(6)
which is just the corrected image centered at y=ccof. The other terms in
Eq. (4) are physically separated in space from the corrected term described
',Prr,TET
Declassified in Part - Sanitized Copy Approved for Release 2012/11/01: CIA-RDP79B00873A000100010186-8
Declassified in Part- Sanitized Copy Approved forRelease2012/11/01 : CIA-RDP79B00873A000100010186-8
UL-VI L. I
'
by Eq. (6) ifcco is made large enough. Although these terms are further
aberrated by the hologram filter, this does not matter since they appear
at different points in space. Thus it appears that the Hologram technique
is one very feasible way to filter out a complex transfer function.
SECRET
Declassified in Part - Sanitized Copy Approved for Release 2012/11/01: CIA-RDP79B00873A000100010186-8
Declassified in Part - Sanitized Copy Approved for Release 2012/11/01: CIA-RDP79B00873A000100010186-8
SECRET
IMAGE RESTORATION
(FIRST-YEAR PROGRAM)
TAB C
15 JUNE 1966
MAJOR GOALS
COMARE TECHNIQUES FOR PRODUCING
HOLOGRAM FILTERS
, 3 4 5 6 7 8 9 10 1 1 12
V
V
V
ESTIMATED COST
Ctri $ Thousa6c1s1
ANAL
YZE THE LIMITING PARAMETERS
OF CHOSEN SYSTEM
DESIGN AND BUILD OPTICAL SYSTEM
FOR PRODUCING FILTERS
EXPERIMENTALLY EVALUATE FILTERS
MAJOR MILESTONES
SECRET
TOTAL F'( _1967 FUNDS
Declassified in Part - Sanitized Copy Approved for Release 2012/11/01 : CIA-RDP79B00873A000100010186-8
25X1
GROUP I?
Excluded km automatic
downgrading and
decloKilicalicn