一种从ERS-1SAR海洋图像中检测船舶航迹的算法

Algorithm to Detect the Ship Wake from ERS-1 SAR Ocean Imagery

简单介绍了船舶航迹的几种模式，分析了单纯运用Ｒａｄｏｎ 变换方法检测航迹的４ 个局限，并针对这些局限提出了一种检测航迹的思想———先进行斑点噪声的抑制，然后用生理视觉的静态边界轮廓系统对图像进行处理，最后应用Ｒａｄｏｎ 变换的方法来突出航迹的线性特征。我们用ＥＲＳ＿１ ＳＡＲ 海洋图像进行了测试，结果表明这一方法比单纯应用Ｒａｄｏｎ 变换的方法更为有效。

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.