作者: Eric Turner , Avideh Zakhor
DOI: 10.1109/ICIP.2012.6467227
关键词: Sampling (statistics) 、 Lidar 、 Radar imaging 、 Computer vision 、 Iterative reconstruction 、 Geometry 、 Boundary (topology) 、 Interpolation 、 Mathematics 、 Surface (mathematics) 、 Artificial intelligence 、 Planar
摘要: In this paper we describe a method for detailed geometry reconstruction of building facades in an urban environment, given 3D point-cloud LiDAR range data. Our approach separates planar faces and interpolates their shape with Moving Least-Squares (MLS) sampling. A is then proposed to reconstruct occluded areas the whereby gaps surface are modeled axis-aligned planes fit gap boundary vertices. This reconstructs unsampled surfaces under assumption that buildings have rectilinear, features. We demonstrate effectiveness our on number facades.