USA flag logo/image

An Official Website of the United States Government

Wide Angle SAR Directed ATR from High Resolution FLIR Imagery

Award Information

Agency:
Department of Defense
Branch:
Air Force
Award ID:
32286
Program Year/Program:
1996 / SBIR
Agency Tracking Number:
32286
Solicitation Year:
N/A
Solicitation Topic Code:
N/A
Solicitation Number:
N/A
Small Business Information
LNK CORP., INC.
6811 Kenilworth Avenue, Suite 306 Riverdale, MD 20737
View profile »
Woman-Owned: No
Minority-Owned: Yes
HUBZone-Owned: No
 
Phase 1
Fiscal Year: 1996
Title: Wide Angle SAR Directed ATR from High Resolution FLIR Imagery
Agency / Branch: DOD / USAF
Contract: N/A
Award Amount: $99,918.00
 

Abstract:

Wide angle sensors provide a large area of coverage to search for targets, though often non-littoral in nature and offer very little in the way of providing details for Automatic Target Recognition (ATR). On the other hand, narrow angle sensors provide the details on the target for ATR, but suffer from a very small area of coverage. To exploit this natural complementarity, in this Phase I, LNK supported by Dr. Rama Chellappa from the University of Maryland, proposes to develop a proof-of-principle sensor fusion ATR system for classification of ground targets from wide angle low resolution and narrow look high resolution sensors. We plan to demonstrate the algorithms using SAR and FLIR imagery. The proposed approach exploits the concept of active vision or directed vision. Our approach involves three major steps of processing. The first step is to detect likely areas of target presence called Focus of Attention (FOA). The second step is to register the SAR image to the FLIR image for extracting the corresponding regions of FOAs detected in SAR using a feature based registration scheme. The third step is to align a selected catalog of BRL-CAD models of targets with the image features in the high resolution FLIR for recognition of the target, using a closed-form line matching technique followed by refinement using a non-linear least squares treatment proposed here. The proposed algorithms will be developed in the Khoros environment.

Principal Investigator:

Dr. S. Raghavan
3019273223

Business Contact:

Small Business Information at Submission:

Lnk Corp, Inc.
6811 Kenilworth Avenue Riverdale, MD 20737

EIN/Tax ID:
DUNS: N/A
Number of Employees:
Woman-Owned: No
Minority-Owned: No
HUBZone-Owned: No