现代电子技术2025，Vol.48Issue(22)：8-14,7.DOI:10.16652/j.issn.1004-373x.2025.22.002
融合LK光流改进的ORB-SLAM2算法研究
Research on ORB-SLAM2 algorithm with improved LK optical flow
摘要
Abstract
In order to solve the problems of poor real-time positioning,insufficient accuracy and inability to densely build maps in the traditional ORB-SLAM2 algorithm,an improved ORB-SLAM2 algorithm with LK optical flow method is proposed.The algorithm is improved based on ORB-SLAM2 architecture.Based on the frame discrimination criterion,the LK optical flow method is added to realize inter-frame accelerated tracking and improve the real-time positioning.The error between the RGB photometric values of the feature points in the image frame and the photometric values of the colour map point cloud is calculated,and used as a constraint to further optimize the pose.An accurate dense point cloud map is constructed.Verification on the TUM dataset and mapping experiments in real scenes show that the average processing time per frame of the algorithm in this paper is 64.99%and 64.75%lower than ORB-SLAM2 and ORB-SLAM3,respectively.The standard error of the positioning accuracy is 31.69%and 29.60%lower than ORB-SLAM2 and ORB-SLAM3 respectively,effectively improving the real-time performance and accuracy of the system.The constructed dense point cloud map is of high value for mobile robot navigation.关键词
改进ORB-SLAM2/LK光流法/特征点提取/帧间加速/光度误差/稠密点云Key words
improved ORB-SLAM2/LK optical flow method/feature point extraction/inter-frame acceleration/photometric error/dense point cloud分类
电子信息工程引用本文复制引用
杨秋宸,王霄,徐凌桦..融合LK光流改进的ORB-SLAM2算法研究[J].现代电子技术,2025,48(22):8-14,7.基金项目
国家自然科学基金资助项目(61861007) （61861007）
国家自然科学基金资助项目(61640014) （61640014）
贵州省科技计划资助项目(黔科合基础-ZK[2021]一般303) （黔科合基础-ZK[2021]一般303）
贵州省科技支撑计划资助项目(黔科合支撑[2022]一般017) （黔科合支撑[2022]一般017）
贵州省科技支撑计划资助项目(黔科合支撑[2022]一般264) （黔科合支撑[2022]一般264）
贵州省科技支撑计划资助项目(黔科合支撑[2023]一般096) （黔科合支撑[2023]一般096）
贵州省科技支撑计划资助项目(黔科合支撑[2023]一般412) （黔科合支撑[2023]一般412）
贵州省科技支撑计划资助项目(黔科合支撑[2023]一般409) （黔科合支撑[2023]一般409）
贵州省教育厅创新群体项目(黔教合KY字[2021]012) （黔教合KY字[2021]012）
中国电力建设股份有限公司科技项目(DJ-ZDXM-2022-44) （DJ-ZDXM-2022-44）
贵大引进人才项目(贵大人基合字(2014)08号) （贵大人基合字(2014）
