一种基于FAST和改进的RANSAC图像匹配方法

An Image Matching Method Based on FAST and Improved RANSAC

为解决双目测量炸点坐标匹配过程中耗时长,高错配性的问题,提出一种基于FAST特征点检测和改进的RANSAC图像匹配方法。改进的RANSAC算法采用自适应阈值方法,具体为将匹配点与其变换模型之间距离的平均值作为算法中阈值选取的依据。实验证明,提出的改进算法可以提高匹配精度和准确率,缩短匹配时间,实现图像的快速匹配。

In order to solve the problem of time-consuming and high mismatch in the process of binocular measurement explosion point coordinate matching, a RANSAC image matching method based on FAST feature point detection and improvement is proposed. The improved RANSAC algorithm adopts the adaptive threshold method, specifically, the average value of the distance between the matching point and its transformation model is used as the basis for the threshold selection in the algorithm. Experiments show that the proposed improved algorithm can improve the matching precision and accuracy, shorten the matching time,and achieve fast image matching.