Target Signatures and Pose Estimation

The paper discusses an important class of defense and commercial applications in the context of Ground Target (Object) Identification, Classification, and Tracking. The data base of target digital signatures is assembled and formed for a full spatial circle (360°) analysis from such sources as High Resolution Radar and Synthetic Aperture Radar. These digital signatures are analyzed from which various spatial as well as frequency (wavelet) characteristics of the targets are formed and interpreted, in order to make good estimate of target pose angle. This angle is key for tracking maneuvering targets. Various statistical measures are obtained from digital signatures to assist in pose angle estimation. We also use certain geometrical considerations to determine an initial pose estimate which is the refined using a variety of correlation coefficients. Expected precision of pose estimate is within few degrees, i.e. within few neighboring target signatures. The paper presents several real life ground target signatures as well as several simulated signatures to illustrate our approach.