Segmentation-based Ship Detection in Harbor for SAR Images

In this paper, we present a novel method to detect the ship in harbor based on image segmentation for SAR imagery. According to the position of the ship in harbor, we divide them into the offshore ship and inshore ship, and different strategies are implemented for ship detection. First, we use the sea-land segmentation method to separate SAR image into land region and sea region, and then extract buffer region according to the coastline. Second, we employ the TS-CFAR detector which has a state-of-the-art performance in multiple-target situations to achieve offshore ship detection. Finally, for the inshore ship, we propose a region-based saliency detection to complete the ship detection. The region-based saliency detection method can tolerate a certain degree of speckle noise. Experimental results show that the proposed method is robust, efficient and can detect different kinds of ship in the harbor.