自适应窗口的SAR立体影像匹配方法  

A stereo SAR matching algorithm with adaptive window

作　　者：王亚超[1,2] 杨书成[2] 赵争[2] 卢丽君[2] 

机构地区：[1]中国矿业大学环境与测绘学院,江苏徐州221116 [2]中国测绘科学研究院,北京100830

出　　处：《测绘科学》2017年第2期135-139,共5页Science of Surveying and Mapping

基　　金：测绘地理信息公益性行业科研专项(201412002);对地观测技术国家测绘地理信息局重点实验室基金项目(K201501);江苏高校优势学科建设工程资助项目

摘　　要：针对SAR立体影像间几何畸变对影像匹配的影响,该文提出一种自适应窗口的SAR立体影像匹配方法。首先引入外部粗分辨率DEM,根据影像参数和DEM高程精度约束匹配搜索范围。然后采用归一化互相关和测度进行辅影像匹配窗口自适应扩展匹配,以保证主辅影像匹配窗口内的灰度信息一致;归一化互相关和值最大时的辅影像匹配窗口为最佳匹配窗口,最佳匹配窗口的中心点为获取的匹配点。最后利用双向一致性约束剔除误匹配点。试验结果表明,本文方法在山地区域能有效地增加正确匹配点数量,提高匹配正确率。Considering the impact of geometric distortions in stereo image pairs, a matching algorithm with adaptive window was proposed. Firstly, image parameters and an external DEM of coarse resolution were utilized to predict the search area. Then the secondary correlation window size was adaptively extended based on summed of normalized cross correlation（SNCC）, so that the primary correlation window and the secondary correlation window covered same region. After that, the optimal secondary correlation window was acquired and the center of the correlation windows was homologous pointed until the value of SNCC raised to maximum. Finally, false match points were eliminated by the left right consistency. The experimental results demonstrated that the proposed algorithm could increase the account of correct homologous points in case of mountainous areas and improve the correct matching rate.

关 键 词：SAR立体 影像匹配 自适应窗口 归一化互相关系数 归一化互相关和 

分 类 号：P237[天文地球—摄影测量与遥感]