基于灰度差分不变量的点特征匹配

本文提出了一种新的基于灰度差分不变量的点特征匹配方法。首先，利用灰度差分不变量获得点集之间的初始匹配；然后，利用初始匹配快速、稳健地估计图象之间的唯一几何约束－对极几何约束；最后，利用对极几何约束改进初始匹配。大量的实际图象实验表明，本文所提出的匹配算法有非常快的运算速度和很高的匹配正确率。

Point Feature Matching Based on Gray-Value Differential Invariants

This paper presents a new point feature matching method based on gray value differential invariants. Firstly, the initial set of matches is estimated using the gray value differential invariants. Secondly, the only available geometric constraint, namely, the epipolar constraint, is exploited robustly using the above initial set of matches. Finally, more accurate matches are eventually found by using the recovered epipolar geometry. The good performance of the proposed algorithm has been justified by extensive experiments on real images. The results show a very fast matching speed and a very high correct matching rate.