RESEARCH AND DEVELOPMENT PROJECT APPROVAL REQUEST FOR AN AUTOMATIC TARGET RECOGNITION STUDY
Document Type:
Collection:
Document Number (FOIA) /ESDN (CREST):
CIA-RDP78B04770A000200020020-7
Release Decision:
RIPPUB
Original Classification:
S
Document Page Count:
3
Document Creation Date:
December 28, 2016
Document Release Date:
October 28, 2004
Sequence Number:
20
Case Number:
Publication Date:
April 1, 1965
Content Type:
MF
File:
Attachment | Size |
---|---|
CIA-RDP78B04770A000200020020-7.pdf | 193.24 KB |
Body:
Approved For Release 2004/11/30: CIA-RDP78B OA00020m
C/D-55-6y
>n:. P7roJeet A vat Request, for
spy
o 1iace with p
it is I"t'gmsun
s:t d rit IAU
25
25
PAUL A. 3
.Aessiattnt itty :
Ixi 1 euc*
Declassification Review by
NGA/DoD
Approved For Release 2004/1.1/30.CIA-RDP78B04770A000200020020-7
Approved For ReleasftZ ff 11/30: CIA-RDP78BO4770A
(
DATE
1 R & D CATALOG FORM
22 March 196
1. PROJECT TITLE/CODE NAME
2. SHORT PROJECT DESCRIPTION
This project is a study to establish the basic design
parameters and performance characteristics of an
X ~utomatic Target Recognitio
operational prototype of an adaptive, automatic, tar-
Study
+ + on q r, PM
3. CONTRACTOR NAME
4. LOCATION OF CONTRACTOR
Falls Church, Virginia
5. CLASS OF CONTRACTOR
6. TYPE OF CONTRACT
Manufacturer
7. FUNDS
8. REQUISITION NO.
9. BUDGET PROJECT NO.
FY 19 $
N/A
NP-S-2
10. EFFECTIVE CONTRACT DATE
11. SECURITY CLASS.
FY 1965 $
(Begin - end)
AA--Confidential
T--Unclassified
FY 1966 $
Jul 1965 - March 1966
W--U
12. RESPONSIBLE DIRECTORATE/OFFICE/PROJECT OFFICER TELEPHONE EXTENSION
DDI NPIC P&DS
13. REQUIREMENT/AUTHORITY
Automated exploitation of reconnaissance imagery to assist PAG, PID, TID
d to be coordinated with T?T)_
an
14. TYPE OF WORK TO BE DONE
Applied Research
15. CATEGORIES OF EFFORT
MAJOR CATEGORY
SUB-CATEGORIES
Interpreter Aids
& St
di
es
u
Special Techniques
16. END ITEM OR SERVICES FROM THIS CONTRACT/IMPROVEMENT OVER CURRENT SYSTEM. EQUIPMENT. ETC.
Final report documenting investigations and establishing final design para-
meters. Interim monthly technical reports will be provided.
17. SUPPORTING OR RELATED CONTRACTS (Agency & Other)/CO DINATION
NPIC currently has a contract with , to build an image prenormali-
zing system. This contract is about 80o complete. The results of this system
will heavily support the proposal design study directed towards a complete, opera
tional system. Many agencies are working in this field; the Air Force has develo
18. DESCRIPTION OF INTELLIGENCE REQUIREMENT AND DETAILED TECHNICAL DESCRIPTION OF PROJECT (Continue on addi-
tional page if required)
This project is directed toward the development of an adaptive, automatic,
target recognition system. It will be designed so as to aid operational inter-
preters by relieving them of the time consuming task of locating and classifying
targets of interest in large volumes of intelligence imagery. The rate of
acquisition of reconnaissance photography is increasing more rapidly than can
be accommodated by a simple increase in exploitation manpower. Therefore, some
reliable means for automating target recognition tasks must be developed.
19. APPROVED BY AND DATE
OFFICE
DEPUTY DIRECTOR
DDCI
2
D00200.02gn2n.
ed
When Filled In) ?
Approved Release 2004/11/3(Z&DP78B'0A000200020020-7
NP-S-?29
R & D Catalog Form continued...
17. (under a contract with an Automatic Target Re=cu;-
nition Device which somewhat parallels approach to the ultimate
problem. GIMRADA is supporting researc or an Automatic Image Classi-
fication System for use in their Rapid Combat Mapping Program; while the
Army Electronics Command is supporting the development of an Automatic
Imagery Screening System for their Tactical Image Interpretation Facility.
Furthermore, the Navy is developing a similar system for A.S.W. applications.
All of these systems have been reviewed and it has been determined that
none duplicate the new system proposed by The number of
current contracts in this field give testimony to the validity of the
operational requirement for this research.
18. Although research has been directed for many years toward the develop-
ment of automated image recognition systems, all of the efforts have been
plagued with several persistent problems. One of the most difficult of
these problems was the fact that adaptive memory computers were very sensi-
tive to image rotation and translation. In other words, before a computer
of the Perceptron type could be taught to recognize an unknown image as
belonging to a class, the unknown image had to be in the same orientation
and in the same part of the field as the images which were used previously
in training the machine. If the unknown images were re-oriented or trans-
lated, the computer probably would not recognize them.
It was decided that until this problem was overcome, there would be
little hope in developing a successful automatic target recognition device.
To attack this problem and others, was awarded a contract to
develop a system to pre-normalize imagery (in rotation and translation)
before the computer was asked to identify it. The resultant system is a
line-integral image scanner and a video signal processor working in-line
with the Conflex I adaptive memory computer. This system is now undergoing
a four month test period during which it will be used to process photography
containing target classes which have many image variables. One type of test
target images being used is U.S. Army tanks which have different orientations
and positions on the image, various levels of resolution and contrast,
various degrees of obscuration, shadow effects, etc. Although these tests
are not complete, early indications are that a system employing this
pre-normalizer and adaptive computer will finally lead to development and
design of an operational Automatic Target Recognition System.
The proposed work in this project is to perform a concerted investigation
of the probl-ems presented by an operational system and to perform a compre-
hensive system design study. If this phase is successful, it is predicted
that an operational Automatic Target Recognition System, useful to the
exploitation community, could be produced in about two years. No pursuit
of the System's development is contemplated unless the investigation phase
is successfully completed.
25
25
25
Approved For Release 2004/1 1/LCf P78B04770A000200020020-7