Identification Algorithm of Extrinsic Parameters Between 3-D Solid-State LiDAR and Mono-Camera

This article introduces an algorithm for identifying the external parameters between a solid-state light detection and ranging (LiDAR) system and a single camera. The algorithm leverages asymmetric 3-D targets to establish saturated constraints and uniquely determines the rotational extrinsic through dual vectors, ensuring that the resulting rotation matrix is orthogonal. Using single-frame images and point cloud data, we derive a closed-form solution for computing the external parameters and employ a global optimizer to enhance their accuracy. In the experimental section, we analyze the impact of various geometric constraints on the observability of external parameters, evaluate the algorithm’s precision, and assess its deployment efficiency on a distributed ARM cluster. Despite the algorithm’s requirement for costly stereo calibration references, it eliminates the need for frequent target or sensor movements, making it particularly suitable for calibrating large-scale equipment.