搭载RGB-D相机轮式机器人误匹配特征点筛除方法  

Method of Removing the Mismatched Feature Points of Wheeled Robot with RGB-D Camera

作　　者：李少波[1] 张伟[1] 孙采鹰[1] 任彦[1] LI Shao-bo;ZHANG Wei;SUN Cai-ying;REN Yan(School of Information Engineering,College of Information Engineering,Inner Mongolia University of Science and Technology,Baotou 014010,China)

机构地区：[1]内蒙古科技大学信息工程学院,内蒙古包头014010

出　　处：《小型微型计算机系统》2020年第11期2359-2364,共6页Journal of Chinese Computer Systems

基　　金：内蒙古自治区自然科学基金项目(2019MS06002)资助.

摘　　要：根据搭载RGB-D相机的平面运动轮式机器人运动特点,本文提出一种基于等高约束条件和直线约束条件的误匹配特征点对筛选方法.本文首先在平面运动机器人相机运动约束条件下,建立了匹配特征点对的等高约束条件和直线约束条件,然后利用该条件筛除了暴力匹配法产生的误匹配点对,取得了高质量的帧间匹配点对.最后,本文将误匹配特征点筛除算法与随机抽样一致性算法相结合,估计了机器人相机的运动.实验证明,与原始随机抽样一致性算法相比,本文提出的特征点筛除算法能够有效提高随机抽样一致性算法的效率,并提高其计算精度.According to the motion characteristics of wheeled planar motion robot carrying RGB-D camera,this paper proposes a mismatching feature points screen out methods based on equal height and straight line constraints.In this paper,firstly,under the motion constraint condition of the planar robot camera,the equal height constraint and the straight line constraint on the matching feature point pairs is established.Then,the condition is used to screen out the mismatching point pairs produced by the violent matching method,and the high-quality matching point pairs between frames is obtained.Finally,the motion of the robot camera is estimated by combining the mismatching feature points screen out algorithm with the random sampling consistency algorithm.Experimental results show that compared with the original random sampling consistency algorithm,the feature point screening out algorithm proposed in this paper can effectively improve the efficiency of the random sampling consistency algorithm and improve its calculation accuracy.

关 键 词：特征点匹配 特征点对筛选 随机抽样一致性 等高约束 直线约束 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]