Building Footprints Extraction of Dense Residential Areas from LiDAR data

Extracting individual buildings and determining their footprints have been extensively studied towards 3D Building reconstruction. Though most previous works show promising results, it is yet a nontrivial task, especially in dense residential areas. This paper discusses a methodology for resolving this issue. The proposed approach starts with separating ground and nonground LiDAR points. In the subsequent step, non-planar points lying on the discontinuity are removed by analyzing the consistency of the normal vectors. The remaining planar points are clustered into a set of individual building blobs. Finally, building footprint for each building is determined by applying alpha-shape and polyline simplification algorithm.