INCREMENTAL MAP REFINEMENT OF BUILDING INFORMATION USING LIDAR POINT CLOUDS

Abstract. For autonomous systems, an accurate and precise map of the environment is importance. Mapping from LiDAR point clouds one promising ways to generate 3D models. However, there are many problems caused by inaccurate data, missing areas, low density points sensor noise. Also, it often not possible or enough a only measurement campaign. In this paper, we propose method incrementally refine several measurements different campaigns represent in hierarchical way with measure indicating uncertainty level detail for objects. The idea thus store all captured information tentative semantics – even when yet complete. Hence, occulated areas presented as well, which can be possibly improved supplemental observation next proposed model framework incremental update evaluated using scans obtained Riegl Mobile System.