Share Email Print
cover

Proceedings Paper

Rapid SAR target modeling through genetic inheritance mechanism
Author(s): Jerzy Bala; Peter W. Pachowicz; Halleh Vafaie
Format Member Price Non-Member Price
PDF $14.40 $18.00

Paper Abstract

The paper presents a methodology and GETP experimental system for rapid SAR target signature generation from limited initial sensory data. The methodology exploits and integrates the following four processes: (1) analysis of initial SAR image signatures and their transformation into higher-level blob representation, (2) blob modeling, (3) genetic inheritance modeling to generate new instances of a target model in blob representation, and (4) synthesis of new SAR signatures from genetically evolved blob data. The GETP system takes several SAR signatures of the target and transforms each signature into more general scattered blob graphs, where each blob represents local energy cluster. A single graph node is describe by blob relative position, confidence, and iconic data. Graph data is forwarded to the genetic modeling process while blob image is stored in a catalog. Genetic inheritance is applied to the initial population of graph data. New graph models of the target are generated and evaluated. Selected graph variations are forwarded to the synthesis process. The synthesis process restores target signature from a given graph and a catalog of blobs. The background is synthesized to complement the signature. Initial experimental results are illustrated with 64 X 32 image sections of a tank.

Paper Details

Date Published: 28 July 1997
PDF: 12 pages
Proc. SPIE 3070, Algorithms for Synthetic Aperture Radar Imagery IV, (28 July 1997); doi: 10.1117/12.281551
Show Author Affiliations
Jerzy Bala, Datamat Systems Research, Inc. (United States)
Peter W. Pachowicz, George Mason Univ. (United States)
Halleh Vafaie, George Mason Univ. (United States)


Published in SPIE Proceedings Vol. 3070:
Algorithms for Synthetic Aperture Radar Imagery IV
Edmund G. Zelnio, Editor(s)

© SPIE. Terms of Use
Back to Top