Highly resolved turntable ISAR signature extraction for ATR

Abstract

By means of data from highly resolved tower-turntable ISAR measurements this paper gives an overview of our work on an ATR process from the raw data acquisition to the final ATR performance evaluation. Main objectives are the radar imaging, the ATR robustness against small changes in the articulation of targets (e.g. military vehicles) and changes in the incidence angle. The recognition process is based on a template matching method. The two-dimensional templates are generated by extracting the most robust scatterers from the RCS image. Once the radar imaging of the target has been performed, an ATR approach is required to extract a maximum of information. In order to investigate the ATR capabilities and the performance, we selected a template matching method. This recognition approach is designed to be tolerant with respect to moderate differences between test and training data. A principal advantage is given by the robustness against artificially removed or produced new scattering centres. The studies were also focused on the determination of robustness against smaller and stronger variations in the articulations of the test objects. Additionally, the dependence on resolution and signal-to-noise ratio (SNR) was explored. Concerning the ATR approach inherently, the results show promising recognition performance.