SAR Segmentation and Recognition Based SCM

A sub aperture cross-correlation magnitude (SCM) algorithm is presented to improve the contrast between targets of our interest and clutters in synthetic aperture radar (SAR) images. Firstly, the SCM processing is applied to the original image. Secondly, the target region is segmented from the SCM image. Thirdly, the feature vector is extracted from the segmented target contour based on the polar mapping (PM). Finally, the correlation coefficient is used to compute the feature vectors of two contours. Experiments on moving and stationary target recognition (MSTAR) data indicate the effectiveness of the target segmentation and recognition with the SCM processing.