The present study looks at the problem of classifying ground targets based on the experimental signatures gathered with a millimeter-wave FMCW radar. The selection of features for robust target classification using circular polarization signatures is considered here. The technique relies on the concept of transient polarization and uses the circularly polarized responses of the targets to define the classification features. Each target can then be decomposed into substructures tracked during the measurements. Once the features of all the substructures are extracted, an artificial neural network is used to classify the signatures.

Date Published: 19 July 2000
PDF: 10 pages
Proc. SPIE 4033, Radar Sensor Technology V, (19 July 2000); doi: 10.1117/12.391846

Published in SPIE Proceedings Vol. 4033:
Radar Sensor Technology V
Robert Trebits; James L. Kurtz, Editor(s)