基于形态成分分析的复杂背景SAR图像舰船尾迹检测  

Ship Wake Detection in SAR Images with Complex Backgrounds Using Morphological Component Analysis

作　　者：杨国铮[1,2] 禹晶[3] 肖创柏[3] 孙卫东[1] 

机构地区：[1]清华大学电子工程系,北京100084 [2]北京市遥感信息研究所处理中心7室,北京100192 [3]北京工业大学计算机学院,北京100124

出　　处：《计算机辅助设计与图形学学报》2016年第10期1662-1671,共10页Journal of Computer-Aided Design & Computer Graphics

基　　金：国家自然科学基金(61501008);首都卫生发展科研专项(2014-2-4025)

摘　　要：SAR图像舰船尾迹检测不仅可用于反演运动舰船的航速航向信息,也有助于发现远小于尾迹的弱小舰船目标.针对现有基于直线检测的舰船尾迹检测方法一般仅适用于简单海况背景SAR图像的问题,提出一种基于形态成分分析的复杂背景SAR图像舰船尾迹检测方法.该方法依据各形态成分只在特定字典下稀疏表示的基本原则,针对舰船尾迹的"线"奇异性及海面纹理的"点"奇异性分别选择不同的基函数构建字典,实现了稀疏意义下舰船尾迹结构成分与海面纹理成分的分离;通过剪切波高频系数重构,实现了舰船尾迹结构成分的增强;最终,使用Radon变换进行舰船尾迹线检测与定位.定性与定量实验结果表明,复杂海况背景下,采用该方法进行舰船尾迹检测的效果明显优于现有其他方法.The detection of ship wakes in SAR images is helpful not only for estimating the speed and the direction of moving ships, but also for finding ship objects which are far smaller than ship wakes. Facing the problem that the existing detection methods based on the line detection can be only used for SAR images under the simple sea background conditions, a novel ship wake detection method for SAR images with the complex backgrounds using the morphological component analysis is proposed. Relied on the basic principle that each component can only be sparsely represented by a specific dictionary, suitable basis functions are selected for the dictionary construction according to the ＂line＂ singularity of ship wakes and the ＂point＂ singularity of sea-surface textures respectively, and the ship wake structure component and the sea-surface texture component are consequently separated in the sense of sparse representation. Then, the enhancement of ship wake structure component is realized through the reconstruction of the high-frequency coefficients of the shearlets transform. And finally, the ship wake lines are detected by using the Radon transform. Both the qualitative and the quantitative experimental results show that, the proposed method conspicuously outperforms the existing methods for the ship wake detection in SAR images, especially under the complex sea background conditions.

关 键 词：SAR图像 舰船尾迹检测 形态成分分析 剪切波变换 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]