Extrinsic Calibration of LiDAR and Camera Using Echo Information

Existing calibration methods for LiDAR and camera extrinsic parameters rely on depth discontinuities to extract edge features from point clouds. However, these methods fail to fully utilize the rich echo information available in modern LiDAR raw data, leading to imprecise calibration results. This paper proposes a novel extrinsic parameter calibration method that leverages LiDAR’s dual-echo capability to enhance calibration accuracy. First, by analyzing the original LiDAR point cloud data format, an algorithm for extracting scene edges using the dual-echo capability of LiDAR is proposed, leveraging echo tag value and reflection intensity in the raw data. Second, a compensation algorithm utilizing the second echo to supplement the data from the first echo is introduced, addressing partially missing points in dual-echo mode. The method achieves an average re-projection error of 2.43 pixels across four scenarios, outperforming similar algorithms.