Implementation and Optimization of LiDAR and Camera Fusion Mapping for Indoor Robots

In this paper, we implement the joint calibration of LiDAR and camera by means of point-surface constraints. The extrinsic parameter is the relative posture of the LiDAR and the camera, providing the basis for point cloud fusion. The odometry uses a fusion odometry based on extended Kalman filtering. Compared to the visual odometry, it combines the data of three sensors of camera, photoelectric encoder and inertial measurement, with higher accuracy and better robustness. The experiment was deployed on the embedded platform Jetson Nano B01. Experimental results show that the system can realize the fusion of two-dimensional LiDAR and RGBD camera, and the accuracy of mapping by the fusion odometry is higher than that of the visual odometry.