Among other things, techniques for identifying background features using LiDAR are described. The techniques include modeling the point cloud information as a sphere and identifying areas that correspond to the respective point clusters of the received LiDAR point cloud information. A graph data structure is generated which contains corner points which correspond to the respective areas of the identified areas, and corner points of the graph data structure are connected based on the adjoining of the respective underlying points and properties of the corresponding areas. Properties of the areas that correspond to the vertices of the subgraph are analyzed, and subgraphs are identified that correspond to a background feature of the physical environment based on the analysis.
展开▼