摘要
针对多源遥感影像间几何变形和灰度差异造成的匹配困难问题,提出一种结合SIFT和边缘信息的影像匹配方法。首先在高斯差分尺度空间进行特征点检测,并采用相位一致性提取可靠的边缘信息;然后结合改进的SIFT和形状上下文对特征点进行描述;最后将欧氏距离和χ2统计作为相似性测度获取同名点。相比于SIFT算法,本文方法可有效地提高匹配正确率,并获得更多的同名点。
In order to address the problem of matching multi-source remote sensing images with geometric distortions and intensity differences, this paper proposes a matching method by combing SIFT and edge information. The feature points are first detected in Difference-of- Gaussian （DOG） scale space, followed by the phase congruency to extract the edge informa- tion. Then the descriptors of the feature points are built by combining improved SIFT and shape context. Finally, we use the Euclidean distance and x2 statistic as the similar metric to determine the correspondences. The experiment results show that the proposed method a- chieves a higher correct matching rate and more correspondences than SIFT matching algo- rithm
出处
《武汉大学学报（信息科学版）》
EI
CSCD
北大核心
2013年第10期1148-1151,1260,共5页
Geomatics and Information Science of Wuhan University
基金
国家973计划资助项目(2012CB719904
2011CB302306
2012CB719901)
中央高校博士研究生自主科研经费资助项目(201121302020002)
