激光雷达与RGB-D相机融合的SLAM建图
SLAM mapping based on fusion of LiDAR and RGB-D camera
付鹏辉 1闫晓磊 1余捷 1于廷海 1叶盛1
作者信息
- 1. 福建省汽车电子与电驱动技术重点实验室,福建 福州 350118
- 折叠
摘要
对二维激光雷达与RGB-D相机联合标定,采用改进的ORB-SLAM2 算法实现稠密的点云地图、八叉树地图、栅格地图的构建.提出了一种将Cartographer算法与改进的ORB-SLAM2 算法融合建图的改进算法.实验结果表明,相比传统的ORB-SLAM2 算法,改进的融合算法在建图过程中障碍物的识别率达到了 96.8%,绝对位姿误差减小了 53.2%,提高了建图的精确性和鲁棒性.
Abstract
For the joint calibration of 2D LiDAR and RGB-D camera,the improved ORB-SLAM2 algorithm was used to construct dense point cloud map,octree map and raster map.An improved mapping algorithm combining Cartographer algorithm with improved ORB-SLAM2 algorithm is proposed.The experimental results show that compared with the traditional ORB-SLAM2 algorithm,the new fusion algorithm can achieve 96.8%obstacle recognition rate and reduce the absolute pose error by 53.2%,which improves the accuracy and ro-bustness of map construction.
关键词
激光雷达/RGB-D相机/ORB-SLAM2/算法/同步定位与建图/多传感器融合Key words
LiDAR/RGB-D camera/ORB-SLAM2 algorithm/SLAM/multi-sensor fusion引用本文复制引用
基金项目
福建工程学院科研启动基金(GY-Z21025)
出版年
2024
NETL
NSTL
万方数据