针对目前图像匹配问题,分析对比了近些年来较为经典的基于点特征的图像匹配算法。归纳了2种较为实用的图像预处理方法: Wallis滤波与灰度均匀化。为提高搜索效率引入了近年来热门的 k⁃dimensional ( KD )树与 Best Bin First ( BBF)结合的匹配方法。总结了除外点的2种前沿方法:双向匹配与Progressive Sample Consensus ( PROSAC)。为基于特征点的图像匹配提供了整体思路。%For the issue of image matching,some recent classical image matching algorithms based on key points are analyzed and compared.Two practical image preprocessing methods,Wallis filter and gray uniformity,are concluded.A popular matching method com⁃bining k⁃dimensional(KD) tree and Best Bin First(BBF) is introduced to improve the search efficiency.Two leading algorithms,bilater⁃al matching and Progressive Sample Consensus(PROSAC),are summarized.As a result,an integrated idea for image matching based on key points is provided.
展开▼