一种改进的ORB特征点匹配算法  

An Algorithm of Improved ORB Feature Point Matching

作　　者：唐管政 唐大全 谷旭平 TANG Guanzheng;TANG Daquan;GU Xuping(Naval Aviation University,Yantai 264001)

机构地区：[1]海军航空大学,烟台264001

出　　处：《舰船电子工程》2021年第6期50-53,共4页Ship Electronic Engineering

摘　　要：ORB算法具有简单快捷的优点,但是原ORB算法存在较多误匹配的现象,针对这一情况,论文对原ORB算法的匹配环节进行改进,提出一种ORB-P算法。通过原ORB特征提取算法提取特征点后,特征匹配环节采用论文提出的基于最近邻比率的特征点前向后向匹配,再使用PROSAC算法剔除误匹配点对,实现精匹配。经过仿真实验证明,改进后的算法具有较强的准确性和实时性,提高了算法的性能,为在复杂场景下的实际应用打下基础。The ORB algorithm has the advantages of simplicity and speed,but the original ORB algorithm has many mismatch⁃es.In view of this situation,this paper improves the matching link of the original ORB algorithm and proposes an ORB-P algorithm.After extracting the feature points through the original ORB feature extraction algorithm,the feature matching link adopts the feature point forward and backward matching based on the nearest neighbor ratio proposed in this paper,and then the PROSAC algorithm is used to eliminate the mismatched point pairs and achieve fine matching.Simulation experiments show that the improved algorithm has strong accuracy and real-time performance,the performance of the algorithm is improved,and the foundation for practical appli⁃cations is laid in complex scenarios.

关 键 词：特征点提取 特征匹配 ORB算法 

分 类 号：TP301.6[自动化与计算机技术—计算机系统结构]