A linear feature mapping framework for adaptive detection and recognition of targets in complex SAR data

A framework for automatic target detection/recognition (ATD/R) system that incorporates modern image processing into classical hypothesis detection theory is presented. The technique is based upon the linear feature map detector. The algorithm is extended to target recognition and applied to the complex SAR data collected by the SRI UHF Ultra-wideband radar which is supported in part by MIT/Lincoln Laboratory. A linear feature mapping such as the discrete cosine transform (DCT) or the wavelet transform (WT) is used to achieve an effective feature representation for target detection and recognition. The effectiveness of the representation is evaluated not only by the number of features, i.e. the dimension of the subspace, needed to represent a target signal for a given mean-square error, but also by the separability of such target features from clutter background and other target classes