利用改进的相位相关算法实现影像亚像素匹配  

Application of Sub-Pixel Matching Based on Enhanced Phase Correlation Algorithm

作　　者：李禄 范大昭[1] 

机构地区：[1]信息工程大学地理空间信息学院,河南郑州450052

出　　处：《测绘科学技术学报》2013年第6期597-600,605,共5页Journal of Geomatics Science and Technology

基　　金：教育部博士后科研基金项目(20080440234)

摘　　要：在影像匹配领域,影像频率域的相位相关技术有着重要的运用。首先简要介绍相位相关基础;然后,接着为了得到亚像素级的匹配精度,提出了一种改进的基于SVD(奇异值分解)相位相关算法。该算法采用Hanning窗来抑制栅栏效应,并采用带权值的相位解缠技术对SVD分解后的特征向量进行处理。最后通过实验证明:提出的改进方法在影像匹配中具有较高的匹配精度。Phase correlation is very important for image matching in frequency domain. Firstly, the theory about phase correlation is introduced and an enhanced phase correlation based on SVD （singular value decomposition） algorithm is applied in order to obtain the precision in sub-pixel level. The algorithm adopts the Hanning windows to reduce barrier effect, and then a weighted-phase unwrap is proposed to process the eigenvectors which has been processed by the SVD algorithm. The experiment result shows that the algorithm has a high precision.

关 键 词：相位相关 亚像素 HANNING窗 SVD分解 相位解缠 

分 类 号：P237[天文地球—摄影测量与遥感]