Colorful Reconstruction from Solid-State-LiDAR and Monocular Version

LiDAR has higher precision depth, longer detection range and stronger robustness in scene light than visual method. In this paper, we fuse a low cost solid-state-LiDAR Livox mid-100 and a camera to generate a colorful and dense 3D reconstruction, which improves the robot's perception and recognition of the environment. By the implement's unique scanning mode, our low cost solid-state-LiDAR can generate denser point clouds than conventional Velodyne series in moving. The LiDAR measures the depth of the scene and the camera records the scene texture. Then paper applies a projection algorithm based on nearest timestamp to generate color reconstruction result. So, our results have the advantages both of LiDAR's accurate depth and version's rich texture. In the experiment, it shows the better display effect reconstruction after fusion. The experimental results demonstrate the robot's ability about perception of the environment.