文档介绍：基于角点检测的自动点匹配算法基于角点检测的自动点匹配算法徐琳曹晓光郁文霞(北京航空航天大学图像处理中心北京 100083 ) 摘要针对多光谱、多传感器遥感图像的自动配准, 本文提出了一个新的基于 Harris 角点检测的高精度自动点匹配算法。该算法充分利用了图像的角点特征以及图像灰度和位置信息, 匹配过程中采用圆形模板进行相似度计算和松弛匹配, 并用全局一致性检测保证匹配的精度。初步实验表明, 本算法对于存在仿射变换的遥感图像可以精确自动匹配,其精度和速度都优于传统的点匹配算法。关键词角点检测点匹配圆形模板一致性检测仿射变换 A Point Matching Algorithm based on Corner Detection Xu Lin Cao Xiaoguang Yu Wenxia ( Image Processing Center, BeiHang University ,BeiJing,100083 ) Abstract For automatic registration of multi-spectral and multi-sensor remote sensing images, a new high accuracy point matching algorithm based on Harris corner detection is described in this paper. It makes full use of graphic point feature, gray-level pixel and location information. In the matching process, circular template is adopted to calculate similarity. Moreover, consistency checking method is used to make sure the matching accuracy... Experimental results on remote sensing image pairs with affine geometric transform showed that the algorithm could do accurate automatic matching and it outperforms traditional point matching algorithms both in accuracy and running speed. Keywords Corner detection Point matching Circular template Consistency checking Affine transform 1引言点匹配是数字图像处理的基本任务之一,在遥感图像处理中被广泛应用, 目前常见的匹配算法主要有: 基于灰度, 基于特征和基于对影像理解和解释等三种算法。基于灰度的匹配直观易懂, 但速度慢、精度低;基于特征的匹配精度高但计算复杂。常用的角点匹配算法有常用的角点匹配算法有Hausdorff 距离法、松弛标记法、确定性退火算法以及迭代最近点算法( ICP ) 角点是像素点在其邻域内的各个方向上灰度变换