Target detection in SAR images using Bayesian Saliency and Morphological attribute profiles

The Synthetic Aperture Radar (SAR) images have become a considerable and vital application in the optical satellite images due to their ability of function in all kinds of weather situations. In synthetic aperture radar (SAR) images Target detection is a demanding task to check presence of targets in images and to exactly locate that target. This paper propose a process for the detection of automatic target in SAR images which has three different steps, that is super pixel segmentation, Morphological attribute profiles and Bayesian Saliency Map. In first stage the SAR input image is then segmented into small regions by super pixels after that the clustering procedure is followed. In the second stage Bayesian saliency map is used to highpoint the clear objects including target and shadow. In the third stage the morphological attribute profile is used to suppress the shadow and highlight the target. The experiment results on MSTAR dataset are compared with some popular existing algorithms (Super pixel CFAR, local CFAR, PR, Global CFAR, MS, BS) which will proves that the proposed mechanism is effective and robust.