LiDAR-Camera Calibration Under Arbitrary Configurations: Observability and Methods

LiDAR-camera calibration is a precondition for many multi-sensor systems that fuse data from LiDAR and camera. However, the constraint from common field of view and the requirement for time synchronization make the calibration a challenging problem. In this paper, we propose a novel LiDAR-camera calibration method aiming to eliminate these two constraints. Specifically, we capture a scan of 3-D LiDAR when both the environment and the sensors are stationary, then move the camera to reconstruct the 3-D environment using the sequentially obtained images. Finally, we align 3-D visual points to the laser scan based on a tightly couple graph optimization method to calculate the extrinsic parameter between LiDAR and camera. Under this design, the configuration of these two sensors is free from the common field-of-view constraint due to the extended view from the moving camera. In addition, we also eliminate the requirement for time synchronization as we only use the single scan of laser data when the sensors are stationary. We theoretically derive the conditions of minimal observability for our method and prove that the accuracy of calibration is improved by collecting more observations from multiple scattered calibration targets. In addition, the proposed method is beneficial to not only plane measurement error-based calibration targets, such as chessboards, but also other point measurement error-based calibration targets, such as boxes and polygonal boards. We validate our method on both simulation and real-world data sets. Experiments show that our method achieves higher accuracy than other comparable methods, which is in accordance with our theoretical analysis.