摘要
针对移动机器人在未知环境中的自身定位问题,提出了一种基于RGB-D相机的移动机器人运动轨迹估计方法.首先,提取当前图像的ORB特征并与关键帧进行特征匹配;然后,采用结合特征匹配质量和深度信息的PROSAC算法对帧间运动进行迭代估计;最后,提取关键帧并利用g2o求解器进行局部优化,得到关键帧位姿的最优估计,进而得到机器人的运动轨迹.实验结果表明:与RANSAC+ICP算法相比,该方法能有效提高移动机器人的定位精度.
Aiming at the problem of localization of mobile robot in unknown environment,a method to estimate the trajectory of robot based on RGB-D camera was proposed.Firstly,ORB(oriented FAST and rotated BRIEF)features were extracted in current frame and matched with key frame.Secondly,combined the quality of feature matching with depth information,the PROSAC(Progressive sample consensus)algorithm was adopted to iteratively estimate interframe motion.Finally,key frame was detected and g2 o(general graph optimization)solver was used to make local optimization to get the best pose estimation of key frame,and then the trajectory of robot was gained.The result shows that this method can effectively improve the positioning accuracy of mobile robot compared with RANSAC+ICP algorithm.
出处
《浙江工业大学学报》
CAS
北大核心
2017年第6期634-638,共5页
Journal of Zhejiang University of Technology
基金
浙江省自然科学基金重点项目(LZ15F030003)