Automatic target recognition of synthetic aperture radar images via gaussian mixture modeling of target outlines

Abstract

A synthetic aperture radar (SAR) target recognition method is proposed in this study based on target outlines. To model the uncertainties of the target outline points caused by noises, extraction errors, etc., the Gaussian mixture model (GMM) is adopted. Afterwards, the distance between two target outlines is measured by the L2 distance between their corresponding GMMs. GMM can robustly describe the possible deformations of the target outline. Therefore, the defined distance measure in this study can better evaluate the inner correlations and differences between different classes using target outlines. By comparing the target outline of the test sample with different template classes based on the designed distance measure, the target label is decided to be the class with the minimum distance. Finally, experiments are conducted on the moving and stationary target acquisition and recognition (MSTAR) dataset for performance evaluation. According to the experimental results, the proposed method achieves a high recognition accuracy of 98.34% in the 10-class recognition problem under the standard operating condition (SOC). In addition, the robustness of the proposed method is also superior over other methods under some typical extended operating conditions (EOCs) including configuration variations, depression angle variance, and noise corruption.