基于相似度曲面二阶差分的景像匹配算法

在分析景像匹配相似度曲面特性的基础上,提出了采用相似度曲面二阶差分为最小作为选取景像匹配点的准则,并进一步给出了利用该准则得到的匹配点的置信度指标. 与传统的景像匹配相似度最大值准则相比,该准则可有效地提高匹配概率(在本仿真实验中,采用修正相关系数(CCF)算法的匹配概率由92.7%提高到98.1%),给出的置信度指标也较好地表示了匹配点的可信程度,并可作为剔除误匹配点的一种参考标准. 

Scene Matching Algorithm Based on Similarity Second-Order Difference

An algorithm was presented, in which the minimum of second-order difference of similarity surface was adopted as the criteria of the selection of scene matching points, on the basis of analyzing the characteristics of scene matching similarity surface. The trustability of the matching point can be further determined by means of above criteria. The experimental results show that this algorithm is effective in improving the matching probability compared with the traditional methods, for example the matching probability improves from 92.7% to 98.1% in this experiment. The confidence levels give a good representation of the matching point evaluation, and can be used as a kind of reference standard for eliminating of mismatching points.