A Robust Target-Based Camera-LiDAR Extrinsic Calibration Method Using Planar Constraints

With the development of autonomous driving, fusing data from different sensors to conduct environmental perception has became an important interest of research. The precise calibration of camera and LiDAR is one of the preliminary demand of camera-LiDAR data fusion. This paper proposes a novel camera-LiDAR extrinsic calibration method to address challenges caused by limited sensor field-of-view (FoV) overlap and unreliable edge-feature extraction. Conventional approaches suffer from LiDAR scanning artifacts (e.g., bleeding points, sparse sampling) and dependency on fully visible calibration targets in point clouds. The proposed method eliminates error-prone edge detection by exploiting planar features, which avoids depth-discontinuity artifacts and relaxes requirements for complete target visibility. The method achieves stable and accurate extrinsic calibration even with partially missing point cloud samples.