A vehicle-borne LiDAR based metro rail detection methodology in metro system is proposed. To begin with, the 2D occupancy grids map in XOY plane is generated from the 3D point cloud and those rail-related points are collected to set up the rail points set. Subsequently, false alarms in the rail points set are excluded in XOZ plane. RANSAC-based line fitting is performed and those outliers of the fitted line are considered as false alarms and thus excluded. Finally, the parameters of rail tracks are estimated with a hybrid method composed of PPHT, RANSAC and CLEAN, fulfilling the task of rail detection. An extra hypothesis points set for RANSAC is created by PPHT and parallelism lines verification to avoid shortening of latter pairs of rail tracks and achieve further removal of false alarms. During the RANSAC based estimation process, the constraint of gauge between pair-wise tracks are concerned and kept, and the chord-length distributions are adopted for overfitting elimination. The CLEAN operation is repeated to seek for multiple pairs of tracks. The proposed methodology has achieved precision of 99.14% and recall of 96.47% on data acquired in field experiments, convincing the effectiveness in the field of rail detection in metro system.
Metro Rail Detection Based on Vehicle-Borne LiDAR
Lect. Notes Electrical Eng.
International Conference on Electrical and Information Technologies for Rail Transportation ; 2021 October 21, 2021 - October 23, 2021
Proceedings of the 5th International Conference on Electrical Engineering and Information Technologies for Rail Transportation (EITRT) 2021 ; Chapter : 6 ; 45-55
2022-02-19
11 pages
Article/Chapter (Book)
Electronic Resource
English
Metro Rail Detection Based on Vehicle-Borne LiDAR
TIBKAT | 2022
|Metro Rail Detection Based on Vehicle-Borne LiDAR
British Library Conference Proceedings | 2022
|Springer Verlag | 2023
|Tram . Light Rail . Metro . Commuter Rail
Online Contents | 2004
Structure-borne Noise Differences of Metro Vehicle Running on Different Tracks
Springer Verlag | 2023
|