Accession Number : ADA303470

Title :   Wavelet Transform Based Target Detection.

Descriptive Note : Final rept.,

Corporate Author : ENVIRONMENTAL RESEARCH INST OF MICHIGAN ANN ARBOR

Personal Author(s) : Subotic, N. S. ; Collins, L. ; Reiley, M. ; Thelen, B. ; Gorman, J.

PDF Url : ADA303470

Report Date : MAY 1995

Pagination or Media Count : 122

Abstract : A novel method of target detection based on a multiple resolution representation of SAR images is described. The signature variation due to interference between scatterers is exploited as resolution is varied. Prominent target scatterers interfere in a characteristic manner as resolution is changed; background noise (clutter) will remain random. Statistical models characterizing the behavior of the multiresolution signature processes for both clutter and cultural objects are derived. A number of detection strategies that exploit these processes are formulated. The detection strategies were applied to simulated SAR scenes containing extended, complex targets in a homogeneous clutter background. Significant detection improvement was demonstrated over pixel by pixel CFAR methods using only the fine resolution signature. These detection strategies were also applied to collected ERIM DCS and Lincoln Laboratory ADTS data. We found that using multiresolution based detection strategies provide a significant increase in detectability of targets.

Descriptors :   *SYNTHETIC APERTURE RADAR, *TARGET DETECTION, MATHEMATICAL MODELS, STRATEGY, RESOLUTION, ELECTROMAGNETIC SCATTERING, VARIATIONS, STATISTICAL ANALYSIS, RADAR CLUTTER, RADAR TARGETS, RADAR SIGNATURES, BACKGROUND NOISE, HOMOGENEITY.

Subject Categories : Active & Passive Radar Detection & Equipment

Distribution Statement : APPROVED FOR PUBLIC RELEASE