Due to the specificity of the rail scene, the research on how to solve online automatic calibration of a 3D LiDAR and a camera is of great practical importance, which enables fused perception of the environment. This paper details the new method of the extrinsic calibration between a 3D LiDAR and a camera depending only on non-calibrating objects in the rail transit. The proposed method converts 2D coordinates of a pixel to 3D coordinates with depth information via the depth estimation of 2D camera. Also, we reduce the count of poses required for calibration to one where the correspondence of 3D lines and planes has been deeply exploited, which enable accurate results to be obtained. Finally, the validity and efficiency of proposed method are verified by the application in actual scenes.
LiDAR-Camera System Automatic Extrinsic Calibration in Rail Transit
2022-10-08
1666993 byte
Aufsatz (Konferenz)
Elektronische Ressource
Englisch