In this thesis, we propose, a Robust data extraction algorithm capable of extracting reliable target features of multiple moving targets of different classes over all channels of an S-Band Doppler Polarimetric Radar PARSAX. The proposed algorithm is capable of generating a time series data by tracking, clusters of detections - representing extended targets using a multi-target tracker modified to track on sequential frames of Range Doppler Maps. The targets considered in this study are Automobiles of different classes (4 wheel drive and above). A performance analysis of the algorithm, for data extraction possibility with respect to target density, has been presented. In addition, the possibility to use the extracted features for Radar Classification has been investigated.