基于水域跟踪的极化SAR图像桥梁检测  

Bridge detection in polarimetric SAR images based on water area tracing

作　　者：刘春[1] 杨健[1] 徐丰 范一大 LIU Chun;YANG Jian;XU Feng;FAN Yida(Department of Electronic Engineering, Tsinghua University, Beijing 100084, China;National Disaster Reduction Center of China, Beijing 100124, China)

机构地区：[1]清华大学电子工程系,北京100084 [2]国家民政部减灾中心,北京100124

出　　处：《清华大学学报（自然科学版）》2017年第12期1303-1309,共7页Journal of Tsinghua University(Science and Technology)

基　　金：国家自然科学基金资助项目(61490693);民用高分重大专项(03-Y20A10-9001-15/16)

摘　　要：从精确的分割和连通桥梁跨越水域角度出发,该文提出一种基于水域跟踪的极化合成孔径雷达(polarimetric synthetic aperture radar,POL-SAR)图像桥梁检测方法。该方法首先使用极化SAR水平集分割方法实现精确的水陆分割,然后通过分支水域扫描及跟踪实现与桥梁检测相关各分支水域的连接,最后提取水域轮廓特征点确定桥梁端点,进而根据桥梁端点确定桥体区域实现桥梁检测。使用新加坡地区和中国海南陵水地区RADARSAT-2极化SAR数据进行实验,实验结果表明:该算法检测正确率高,且在大场景范围内实现了极短桥梁检测。Bridges interfere with automatic segmentation and linking of waterways crossed by bridges. This paper describes an automatic bridge detection method using polarimetric synthetic aperture radar （SAR） based on tracing of known water areas. The water areas are first accurately extracted using a level set segmentation algorithm from polarimetric SAR images. All the water branches related to the bridge are then linked by the scanning and the tracing of the water areas. The feature points on the water branch contours are extracted to identify, the bridge end points. The bridge bodies are then developed from the end points. The algorithm was tested using polarimetric SAR images from RADARSAT-2 over Singapore and Lingshui, China. The results show that the detection rate of this algorithm is very high. Even small bridges in large scenes are correctly identified by the algorithm.

关 键 词：合成孔径雷达 桥梁检测 水域跟踪 

分 类 号：TN957.52[电子电信—信号与信息处理]