基于优化ORB算法的图像角点特征匹配方法  

An Image Corner Features Matching Method Based on Optimized ORB Algorithm

作　　者：朱标 ZHU Biao(AVIC Huadong Photoelectric Co., Ltd.,Anhui Wuhu 241002, China;Key Laboratory of Modern Display Technology,Anhui Wuhu 241002, China;National Special Display Engineering Research Center, Anhui Wuhu 241002, China;National Engineering Laboratory of Special Display Technology,Anhui Wuhu 241002, China)

机构地区：[1]中航华东光电有限公司,安徽芜湖241002 [2]安徽省现代显示技术重点实验室,安徽芜湖241002 [3]国家特种显示工程技术研究中心,安徽芜湖241002 [4]特种显示国家工程实验室,安徽芜湖241002

出　　处：《重庆工商大学学报（自然科学版）》2021年第4期42-51,共10页Journal of Chongqing Technology and Business University:Natural Science Edition

基　　金：安徽省科技重大专项项目(17030901053)资助.

摘　　要：针对传统ORB算法的图像角点特征匹配精度不高的问题,提出基于优化ORB算法的图像角点特征匹配方法;首先使用Shi-Tomasi算法检测图像角点特征,然后使用BRIEF和SURF相融合算法生成图像角点特征双描述子序列并使用随机投影原理进行降维,最后使用优化的匹配算法进行匹配,简称Shi-Tomasi-SURFORB算法,仿真实验通过统计角点特征数、角点特征匹配数、角点特征正确匹配数、角点特征匹配精度、角点特征匹配精度率、图像匹配时间和图像匹配时间率共7个指标进行分析;分析结果表明:Shi-Tomasi-SURFORB算法与传统ORB算法相比,在图像角点特征匹配时间方面提升了9.46%,但在图像角点特征匹配精度方面提升了8.88%,为图像角点特征匹配提供了一种更加均衡的解决方案。Aiming at the problem that the image corner features matching accuracy of the traditional ORB algorithm is not high,an image corner features matching method based on optimized ORB algorithm is proposed.Firstly,this paper uses the Shi-Tomasi algorithm to detect the image corner features,then uses the fusion algorithm of BRIEF and SURF to generate the double-descriptor sequences of the image corner features,and reduces dimension of the double-descriptor sequences by the random projection principle,finally,the optimized matching algorithm is used to match,referred to as the Shi-Tomasi-SURFORB algorithm.The simulation experiment analyzes seven indicators,including the number of corner features,the number of corner features matching,the number of correct corner features matching,the accuracy of corner features matching,the corner feature matching accuracy ratio,the image matching time,and the image matching time rate,the analysis results show that compared with the traditional ORB algorithm,the Shi-Tomasi-SURFORB algorithm improves the image corner features matching time by 9.46%,but improves the image corner features matching accuracy by 8.88%,which provides a more balanced solution for image corner features matching.

关 键 词：ORB算法 Shi-Tomasi算法 BRIEF算法 SURF算法 特征匹配 

分 类 号：TP394[自动化与计算机技术—计算机应用技术]