基于极线局部校正的特征匹配方法

针对利用多视图像进行目标三维测量、结构重建时在极线约束下对直接区域灰度相关进行同名特征匹配常常失效的问题,提出了一种基于极线局部校正的特征匹配算法。介绍了极线约束匹配的原理,分析了相关方法在极线约束匹配中的缺陷以及在多种像机位姿配置下的图像特征间的约束关系,在此基础上提出了一种极线局部区域校正的方法,对待匹配区域进行校正使自动相关匹配能有效执行,结合最小二乘匹配得到了高精度的匹配结果。实验结果证明了新算法的有效性,大大提高了自动匹配的可靠性、速度和精度。

Feature matching algorithm based on rectification of epipolar-line region

In the process of 3D measurement and structure reconstruction from multi-views,the direct area-correlation method based on epipolar constraint works inefficiently to match targets.To solve this problem,a feature matching algorithm based on rectification of epipolar-line region is proposed.The theory of epipolar constraint is introduced,the correlation matching method and the corresponding feature property of image pair are analyzed,the algorithm is introduced.The algorithm rectifies the regions of the corresponding epipolar-lines which include the feature points to be matched through rotating.Matching by correlating on the rectified regions works automatically and gives accurately correspondence throuth least square image matching.Experiments with real images show that the algorithm improves the robustness and the accuracy of feature matching method under epipolar constraint.