基于空间子区一致性的异源图像匹配方法  

Multi-sensor Image Matching Based on Spatial Sub-area Congruency

作　　者：李壮[1] 杨夏[1] 雷志辉[1] 

机构地区：[1]国防科技大学航天与材料工程学院,湖南长沙410073

出　　处：《国防科技大学学报》2011年第1期31-34,共4页Journal of National University of Defense Technology

基　　金：国家高技术研究发展计划资助项目(2007AA12Z121)

摘　　要：提出一种基于空间子区一致性的异源图像匹配方法。该方法中,将图像划分为许多大小相同的子区;分别计算邻近子区间的灰度直方图距离,得到一种图像结构特征描述———似邻矩阵;利用似邻矩阵计算图像间的子区一致性;采用改进的搜索策略,快速获得匹配结果。实验结果表明,该方法在正确率和计算时间方面均优于传统方法。A novel multi-sensor image matching method based on spatial sub-area congruency is proposed.In this method,images were segmented into sub-areas with same size.The structure feature represented by matrix of surrounding similarity was obtained by calculating the similarity of intensity histogram of each pair of adjacent sub-areas,and was used to compute spatial sub-area congruency.Through employing optimized searching algorithm,the matching result was acquired.Experimental results show that the proposed method can match multi-sensor images stably and fast and outperforms conventional methods.

关 键 词：图像处理 图像匹配 似邻矩阵 子区一致性 异源图像 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]