一种利用相位信息的异源遥感影像配准方法  

Multi-source Remote Sensing Images Registration Method Based on Phase Information

作　　者：舒建英 白兰东 李中华 叶沅鑫[2] 

机构地区：[1]成都市国土规划地籍事务中心,四川成都610000 [2]西南交通大学地球科学与环境工程学院,四川成都610031

出　　处：《地理空间信息》2014年第3期81-83,7,共3页Geospatial Information

基　　金：国家973计划资助项目(2012CB719901)

摘　　要：提出一种利用相位信息的异源遥感影像配准方法。首先对两幅影像进行相位一致性变换以消除影像间的灰度和对比度差异,并利用相位一致性最小矩在参考影像上提取特征点,然后以相关系数作为相似性测度,在输入影像上采用模板匹配策略获取同名点,最后采用投影变换实现影像配准。实验表明,该方法对于异源遥感影像具有较强的适应性,且配准精度较高。A registration method using phase information for multisource remote sensing images was proposed in this paper. Firstly, we calculated the phase congruencies of images to eliminate the gray and contrast dif erences. Minimum moment of phase congruency was used to select feature points in the reference image. Then, we employed the correlation coefficient as the similarity metric to determine the correspondences in the input image by the template-matching scheme. Finally, the registration was achieved using the projective transform. The experiment results demonstrate that the proposed method is suitable for multi-source remote sensing images, and achieves high registration accuracy.

关 键 词：影像配准 相位一致性 相关系数 精度分析 

分 类 号：P237.4[天文地球—摄影测量与遥感]