Multi-Spectral Fusion Techniques

Period of Performance: 03/12/1997 - 03/15/1999

$750K

Phase 2 SBIR

Recipient Firm

Data Fusion Corp.
10190 Bannock Street Suite 246
Northglenn, CO 80260
Principal Investigator

Research Topics

Abstract

Data Fusion Corporation (DFC), together with Lockheed-Martin Missiles & Space and Oasis Research Center, will jointly develop algorithms and their software implementations for studying radiometric invariants for thermal images, hyperspectral images and high resolution radar data for ATR application. In addition, we propose to continue the statistical analysis from the Phase I research for computing and updating probablilities of weak evidence derived from invariants for ATR. The proposal inlude these features: - Signal Processing for thermal invariants -Compensation for atmospheric and noise effects -Canonical correlation for quantization -Canonical correlation for performance analysis - Hyperspectral invariants -proportionality-independent invariants -illumination-independent invariants -Range image creation, interpolation, and segmentation using geometric invariants, thermal invariants, and concepts from differential geometry -HRR invariants -Invariant evidential accural for ATR