摘要
针对星载可见光与SAR图像配准中存在辐射差异大造成匹配精度低的问题,提出一种多特征匹配的可见光与SAR图像配准算法。基于多级配准策略,首先在低分辨率尺度下利用基于SIFT特征的方法求取初始变换,然后在原始分辨率下借助初始变换的引导在SAR图像上提取直线、区域等结构特征并完成与可见光图像相应特征的精确匹配。以3mTerraSAR卫星和1mIkonos光学卫星数据为例,实验结果表明该算法能够有效提升图像数据配准精度,具备良好的工程化应用基础。
To solve the registration problem between optical and SAR images caused by obvious radiometric differences between them,this paper presents a multi-feature matching algorithm for registration between optical and SAR images.In the multi-stage registration strategy,a method based on SIFT features is firstly employed under the low resolution scale to obtain the initial transformation parameters,and then structure characteristics of the SAR image,e.g.,lines,regions,etc.,are extracted by the use of the initial transformation parameters under the original resolution.Finally,the precise matching with the corresponding feature of optical image is completed.Taking TerraSAR 1 msatellite and Ikonos 3 moptical satellite data as an example,the experimental results illustrate that this algorithm can effectively improve the registration accuracy,and has a good foundation for the engineering application.
作者
许斌
雷斌
孙韬
卢晓军
XU Bin;LEI Bin;SUN Tao;LU Xiaojun(Institute of Electronics,Chinese Academy of Sciences,Beijing 100190,China;University of Chinese Academy of Sciences,Beijing 100049,China;Key Laboratory of Technology in Geo-spatial Information Processing and Application System,Institute of Electronics,Chinese Academy of Sciences,Beijing 100190,China;China International Engineering Consulting Corporation,Beijing 100048,China)
出处
《遥感信息》
CSCD
北大核心
2018年第3期85-90,共6页
Remote Sensing Information
关键词
图像配准
遥感图像
合成孔径雷达
多特征
初始变换引导
image registration
remote sensing image
synthetic aperture radar
multi-feature
initial transformation guidance
