Synthetic Aperture Radar 3D Feature Extraction for Arbitrary Flight Paths

Abstract

We propose an algorithm for extracting multiple geometric scattering features from synthetic aperture radar (SAR) phase history collected over arbitrary, 3D monostatic or bistatic apertures. The algorithm input is complex-valued phase history; the output is a list of features and corresponding parameter estimates. We fit to the data parametric models for six canonical features. The feature extraction problem includes model order selection, shape classification, and parameter estimation. Examples include densely-sampled and sparse apertures for monostatic and bistatic scenarios.