Extracting Global 3D Scattering Center Model of Radar Target from Multiple HRR Profiles

Abstract

Global 3D scattering center model represents the radar target in optical region by a set of point scatterers in three dimensional space with spatial scattering patterns. Such a model is useful in scattering analysis, data compression and target recognition. R. Bhalla et al. have proposed to extract such a model from multiple 3D radar images at viewing angles on a mesh grid covering the space extent of interest. Here we propose another method to extract the model from multiple 1D radar images, i.e., high range resolution profiles. This alleviates the tremendous task of collecting enough data for 3D imaging at each viewing angle. We explore the fact that the projected 1D locations of a fixed scattering center at different viewing angles are on the same sphere and we devise a 1D-3D scatter map (OTSM) to visualize those spheres. Though OTSM gives a straightforward demonstration of the 3D scattering center model, we devise a procedure to extract the model automatically, which makes use of algorithms such as Hough transform and morphologic operation. The model we extracted contains the stable scattering centers of the target and is valid in a large space extent. It can be used to restore the target signatures such as RCS and 1D to 3D radar images at any viewing angles in this space extent and at any frequencies in or near the band of the original data. Examples using both simple and complex targets demonstrate the validity of this method.