Accession Number : ADA332033
Title : New Models and Fast Algorithms for Natural and Urban Clutter with Applications
Descriptive Note : Final rept. Aug 95-Jan 97,
Corporate Author : BROWN UNIV PROVIDENCE RI
Personal Author(s) : Cooper, David B.
PDF Url : ADA332033
Report Date : JUL 1997
Pagination or Media Count : 33
Abstract : To address the Automatic Target Detection/Recognition (ATD/R) community's long term goal of 'revolutionizing wide-area imagery analysis', the research presented in this report focused on two enabling technologies: (1) clutter models based on Synthetic Aperture Radar (SAR) in urban environments; and (2) fast algorithms for Markov Random Fields. The former investigated the nature of building signatures in SAR imagery, and saw the development of a building detector. Buildings contribute a large number of false target detections, and targets stationed in close proximity to buildings can be missed using conventional analysis methods. The latter effort revisited signal processing to make theoretical headway in developing reduced computational cost techniques for estimating and using stochastic target and clutter models based on Markov Random Fields. These algorithms have broad application to a large variety of ATD/R concerns.
Descriptors : *IMAGE PROCESSING, *MARKOV PROCESSES, *COST BENEFIT ANALYSIS, ALGORITHMS, SIGNAL PROCESSING, COMPUTATIONS, STOCHASTIC PROCESSES, RANDOM VARIABLES, TARGET RECOGNITION, SYNTHETIC APERTURE RADAR, TARGETS, RADAR CLUTTER, FALSE TARGETS, URBAN AREAS.
Subject Categories : Economics and Cost Analysis
Statistics and Probability
Distribution Statement : APPROVED FOR PUBLIC RELEASE