Advanced HSI Change Detection Algorithms for Exploitation of Space-Based Imagery

Period of Performance: 02/03/2006 - 12/14/2009

$2.56MM

Phase 2 SBIR

Recipient Firm

Space Computer Corp.
12121 Wilshire Boulevard
Los Angeles, CA 90025
Principal Investigator

Abstract

This proposed Phase II effort will capitalize on recent advances in precision registration and coherent change processing of hyperspectral imagery (HSI) to significantly enhance the military s ability to detect, identify and track changes on the ground from spacecraft or high-altitude aircraft. Existing HSI processing approaches that rely solely on spatial and spectral information fall short of demanding DoD needs because of the complexity of the ground clutter environment and practical sensor limitations on simultaneous acquisition of high-resolution spatial, spectral and temporal scene information at long range. The Phase I SBIR showed the feasibility of obtaining large improvements in change target sensitivity, relative to conventional single-look approaches, using multi-look HSI observations at spatial and spectral resolutions commensurate with planned space-based sensors. In Phase II we propose to further develop the core algorithms for registering, predicting and comparing spectral image observables in a robust change analysis framework, implement the algorithms in a prototype software tool, and apply them to current and near-term sources of satellite HSI. These improved processing techniques will significantly enhance HSI and MSI sensor utility for tactical target detection, surveillance and monitoring missions, particularly for low-observable subpixel targets with sizes comparable to or smaller than the sensor s spatial resolution.