基于FAST和SURF的特征点快速匹配算法  

Fast feature point matching algorithm on FAST and SURF

作　　者：产叶林 胡新平[1] CHAN Ye-lin;HU Xin-ping(School of Electronic Information,Nantong University,Nantong 226019,China)

机构地区：[1]南通大学电子信息学院

出　　处：《计算机工程与设计》2019年第12期3500-3504,共5页Computer Engineering and Design

摘　　要：为解决传统印刷电路板(PCB)图像配准过程中匹配耗时和错配率较高的问题,提出一种基于FAST-SURF的特征点匹配优化算法。利用FAST算法快速提取特征点,利用SURF的64维描述子进行准确的特征描述,在匹配阶段使用K-Means算法优化匹配结果,通过RANSAC算法进行一致性检查,消除误匹配点。实验结果表明,与传统的SURFRANSAC算法相比,该算法提高了匹配正确率,减少了匹配时间,实现了PCB图像特征的快速匹配。To solve the problem of time-consuming matching and high mismatch rate in traditional printed circuit board(PCB)image registration,a feature point matching optimization algorithm based on FAST-SURF was proposed.The FAST algorithm was used to extract feature points quickly,the 64-dimensional descriptor of SURF was used to describe the features accurately,the K-Means algorithm was used to optimize the matching results in the matching stage,and the consistency check was carried out using RANSAC algorithm to eliminate the mismatching points.Experimental results show that,compared with the traditional SURF-RANSAC algorithm,the proposed algorithm improves the matching accuracy and reduces the matching time,achieving fast matching of PCB image features.

关 键 词：特征点匹配 FAST检测 加速鲁棒特征 K-MEANS算法 随机采样一致性 

分 类 号：TP391.4[自动化与计算机技术—计算机应用技术]