The content of this article comes from the "Surveying and Mapping Bulletin" No. 5, 2024, drawing review number: GS Jing (2024) No. 0777
3D scene reconstruction system and algorithm based on binocular vision and single-line lidar
Zhong Leisheng, Xia Hui, Chen Jialin
Unit 92728, Shanghai 200436
Keywords: binocular vision, lidar, 3D reconstruction
Citation format: Zhong Leisheng, Xia Hui, Chen Jialin. 3D Scene Reconstruction System and Algorithm Based on Binocular Vision and Single-line LiDAR[J]. Bulletin of Surveying and Mapping, 2024(5): 48-52,59.doi: 10.13474/j.cnki.11-2246.2024.0509
Abstract:Binocular vision and lidar are two effective methods to achieve 3D scene reconstruction, but they both have their own limitations. It is important to fuse data from vision sensors and laser sensors to overcome their respective shortcomings. In view of the uniqueness of single-line rotating LiDAR equipment, a modular Binocular Vision-LiDAR SLAM (Stereo-LiDAR SLAM) system and algorithm based on binocular images and LiDAR data are proposed. In this method, the binocular vision information is used to eliminate the distortion of the lidar point cloud, and the binocular vision localization (VO) module is supported to provide the initial pose estimation for the whole system. Then, the LIDAR SLAM (L-SLAM) module, which is independent of the VO module, optimizes the pose parameters to obtain high-precision 3D scene reconstruction results. Experiments show that the system and algorithm designed in this paper can effectively improve the accuracy and environmental adaptability of large-scale low-cost 3D scene reconstruction.
About author:ZHONG Leisheng (1991—), male, Ph.D., research direction is computer vision and 3D reconstruction. E-mail:[email protected]
First instance: Yang Ruifang review: Song Qifan
Final Judge: Jin Jun
information