摘要
同一场景下的合成孔径雷达(SAR)图像的灰度特性由于相关噪声的影响及成像条件不同可能存在很大差异,使得单纯基于边缘特征或灰度信息的方法难以胜任SAR图像配准工作。根据SAR图像的特点,提出一种典型地物边缘形状信息与局部灰度统计信息相结合的基于特征的图像配准方法,弥补了仅利用边缘特征或灰度信息的方法在SAR图像配准中的不足。给出了本方法用于Radarsat图像上的实验结果。
Synthetic Aperture Radar(SAR) images of same scene may have obvious difference in gray\|level characteristic because of correlation noise and different imaging conditions. Method based only on edge feature or gray\|level information doesn't adapt to SAR image registration. This paper describes a new feature\|based approach using edge information of typical objects combined with local gray\|level statistical information to image\|to\|image registration. First, local area with typical object is selected instead of the whole image registration to reduce computation time. In each area, edge feature of the typical object is extracted and initial feature point set is decided according as shape information of the points on edges. Then, a principle combined with edge shape and gray\|level information is used on feature point matching to reduce the effect which the accuracy of edge extraction may add on the accuracy of registration. At last, a root mean square error is computed. Experimental results with pixel level accuracy on Radarsat imagery proves the validity of this approach.
出处
《遥感技术与应用》
CSCD
2003年第3期159-163,共5页
Remote Sensing Technology and Application
基金
微波成像技术国家重点实验室基金资助项目(编号:51442030101ZK1301)
关键词
图像配准
特征提取
合成孔径雷达
Image registration, Feature extraction, Synthetic aperture radar