Algorithm to Detect the Ship Wake from ERS-1 SAR Ocean Imagery
一种从ERS—1 SAR海洋图像中检测船舶航迹的算法

Detection of the ship wake from SAR imagery is very significant. Most of the methods detecting ship wakes from SAR imagery are based on Radon transform appending condition test using prior knowledge. In the paper we introduce the patterns of three main kinds of ship wakes. Then we analyse four limitations of the method that applies Radon transform to detect ship wake in SAR ocean images. In order to improve the limitations we describe another method to detect V_shape ship wake. At first, we filter the SAR images to suppress the speckle noise, then, use Static Boundary Contour System based on biological vision to process the filtered images. Finally Radon transform is applied to enhance the ship wake. We describe two parts of the Static Boundary Contour System in the paper. The first part is Static Oriented Contrast which is used to determine the local good orient of the contrast difference in the input, and the second pasrt is Cooperative_Competitive Loop which is used to sharpen, group and select the edge. We test the algorithm by using ERS_1 SAR coean images. The result deriving from the new algorithm is better than that from simple Radon transform.