Enhanced Low-resolution LiDAR-Camera Calibration Via Depth Interpolation and Supervised Contrastive Learning
This work improves calibration accuracy for applications using low-resolution LiDAR, such as autonomous vehicles or robotics, but is incremental as it builds on existing techniques.
The paper tackles low-resolution LiDAR-camera calibration by addressing sparsity and noise in point clouds, achieving average mean absolute rotation and translation errors of 0.15 cm and 0.33° on 32-channel LiDAR data, significantly outperforming reference methods.
Motivated by the increasing application of low-resolution LiDAR recently, we target the problem of low-resolution LiDAR-camera calibration in this work. The main challenges are two-fold: sparsity and noise in point clouds. To address the problem, we propose to apply depth interpolation to increase the point density and supervised contrastive learning to learn noise-resistant features. The experiments on RELLIS-3D demonstrate that our approach achieves an average mean absolute rotation/translation errors of 0.15cm/0.33\textdegree on 32-channel LiDAR point cloud data, which significantly outperforms all reference methods.