基于稀疏光流法的改进ORB特征匹配算法  

Improved ORB Matching Algorithm Based on Sparse Optical Flow Method

作　　者：陶卓 黄卫华 章政[1,2,3] 姚艺 何佳乐[1] TAO Zhuo;HUANG Wei-hua;ZHANG Zheng;YAO Yi;HE Jia-le(Institute of Robotics and Intelligent Systems,Wuhan University of Science and Technology,Wuhan 430081,China;Engineering Research Center for Metallurgical Automation and Measurement Technology of Ministry of Education,Wuhan University of Science and Technology,Wuhan 430081,China;School of Information Science and Engineering,Wuhan University of Science and Technology,Wuhan 430081,China)

机构地区：[1]武汉科技大学机器人与智能系统研究院,武汉430081 [2]武汉科技大学冶金自动化与检测技术教育部工程研究中心,武汉430081 [3]武汉科技大学信息科学与工程学院,武汉430081

出　　处：《组合机床与自动化加工技术》2021年第12期1-5,共5页Modular Machine Tool & Automatic Manufacturing Technique

基　　金：国家自然科学基金(61773298);冶金自动化与检测技术教育部工程研究中心开放基金资助项目(MADT201603);2017年度武汉科技大学国防预研基金项目(GF201706)。

摘　　要：考虑到无人机运动、机载摄像头采集效果、信号传输以及环境变化等因素,针对传统ORB特征匹配算法存在匹配精度低、鲁棒性差等问题,提出了一种基于稀疏光流法的改进ORB特征匹配算法。首先,采用高斯滤波对无人机采集的图像进行预处理,并采用Harris算子剔除特征点中伪角点,提高图像特征点的质量;其次,采用K-means++算法实现图像特征点的聚类处理,由此减少暴力匹配的计算量;在此基础上,将稀疏光流法引入ORB特征匹配算法,计算特征点的运动向量并根据估计特征点在待匹配图像中的二维坐标,实现剔除偏离聚类中心较远的特征点匹配对,再经过随机抽样一致性算法优化后最终特征匹配结果。实验结果表明了所设计的改进ORB特征匹配算法可行性和有效性。Considering the UAV motion,airborne camera acquisition effect,signal transmission and environmental changes,an improved ORB feature matching algorithm based on sparse optical flow method is proposed to solve the problems of low matching accuracy and poor robustness of traditional ORB algorithm.Firstly,Gaussian filter is used to preprocess the image collected by UAV,and Harris operator is used to eliminate the false corners in the feature points to improve the quality of the image feature points.Then,K-means++clustering algorithm is used to realize the clustering of image feature points,so as to reduce the calculation of violent matching.On this basis,the sparse optical flow method is introduced into ORB feature matching algorithm to complete the calculation of the motion vector of the feature points.According to the estimated two-dimensional coordinates of the feature points in the image to be matched,the feature points far away from the cluster center are eliminated,and the final feature matching result is optimized by random sampling consistency algorithm.Finally,the experimental results show the feasibility and effectiveness of the improved ORB feature matching algorithm.

关 键 词：ORB算法 稀疏光流法 机载视觉导航系统 K-means++ 无人机 

分 类 号：TH166[机械工程—机械制造及自动化] TG506[金属学及工艺—金属切削加工及机床]