Octree-Based Level Progressive Point Cloud Registration Framework

Among many 2D and 3D SLAM algorithms, scan matching and map merging are extremely critical steps. Among these steps, the registration of point cloud plays the most important role, which directly affects the localization and mapping accuracy. Aiming at the conflict between accuracy and efficiency caused by the huge amount of data in dense point cloud registration, we propose a Level Progressive Registration framework based on Octree structure, which is used for the registration processing of dense point clouds, and has been verified by experiments. The aim of this work is to accelerate the process of dense point cloud registration and utilize the organizational role of the Octree in space to handle point cloud registration issues in a more robust manner.