New Models and Fast Algorithms for Natural and Urban Clutter with Applications
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.
Document Details
- Document Type
- Technical Report
- Publication Date
- Jul 01, 1997
- Accession Number
- ADA332033
Entities
People
- David B. Cooper
Organizations
- Brown University