Detection of Obstacles in Tunnel Based on Vehicle-borne LiDAR

Massive point clouds introduced by tunnel wall reflection can easily cause false alarms in LiDAR-based detection of obstacles in tunnel environment. A vehicle-borne LiDAR based obstacle-in-tunnel detection methodology is proposed in this paper. Firstly, a strategy of removing background point cloud...

Full description

Saved in:
Bibliographic Details
Main Authors: ZENG Xiang, JIANG Guotao, BAO Jiyu, LIU Bangfan, XIAO Zhihong
Format: Article
Language:zho
Published: Editorial Office of Control and Information Technology 2021-01-01
Series:Kongzhi Yu Xinxi Jishu
Subjects:
Online Access:http://ctet.csrzic.com/thesisDetails#10.13889/j.issn.2096-5427.2021.01.100
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:Massive point clouds introduced by tunnel wall reflection can easily cause false alarms in LiDAR-based detection of obstacles in tunnel environment. A vehicle-borne LiDAR based obstacle-in-tunnel detection methodology is proposed in this paper. Firstly, a strategy of removing background point cloud is designed. 2D grid map is generated from 3D point cloud, and grids corresponding to the tunnel boundary or the ground are labeled respectively. Based on the Euclidean clustering algorithm, the point cloud corresponding to the tunnel boundary is extracted. With the estimation of the parameters of the boundary curves, point cloud corresponding to the tunnel boundary is further removed. Similarily, the point cloud corresponding to the ground is also removed based on the estimation of parameters of a space plane. Subsequently, the obstacles are extracted independently from the remaining point cloud by Euclidean clustering, followed by the estimation of the position and dimensions of all those obstacles. Finally, the obstacles tracking is achieved by means of the global nearest neighbor algorithm with improved distance metric and the Kalman filter, and the track of all obstacles is updated by a customized life state transition strategy. Experimental results show that the proposed method can eliminate the interference of background point cloud effectively and yield stable results of obstacles identification and tracking.
ISSN:2096-5427