Science of Land Target Spectral Signatures
Abstract
This MURI program progress report covers the fifteen months from 1 August 2007 to 30 September 2008. The overall purpose of this program is a study of the science underlying the signatures generated by land targets (both natural and manmade). Specifically, the objective is to understand the dependency of hyperspectral and hyperspectral/polarimetric signatures on variations in environmental conditions. An additional objective involves algorithm development and focuses on the identification of new target discriminants and the evaluation of their utility for target detection and sensor fusion. The phenomenology research continued to focus on spectroscopic soil measurements, optical property analyses, field data analysis, physics-based signature modeling, and synthetic scene generation for concealed targets. The algorithm work extended the signature-based detection research and conducted investigations into patterned based detection methods. Additional algorithm work exploiting local spatial-spectral background characteristics was conducted. Two endmember selection algorithms were developed based on a sparsity promoting approach. Fusion research continued to investigate methods for fusing SAR and hyperspectral data through the Choquet Integral approach. Target detection performance improved using this data fusion approach. Clutter complexity research continued to investigate metrics to tie image complexity to detection algorithm performance and continued work on methods for generating diverse sample images.
Document Details
- Document Type
- Technical Report
- Publication Date
- Apr 03, 2013
- Accession Number
- ADA575755
Entities
People
- J. M. Cathcart
Organizations
- Georgia Tech