A Methodology for Automated Segmentation and Reconstruction of Urban 3-D Buildings from ALS Point Clouds

Abstract

In this paper, a methodology which allows automated and efficient reconstruction of three-dimensional (3-D) geometric building models from an Airborne Laser Scanning (ALS) point cloud is introduced and its performance is analyzed and evaluated. The proposed method avoids abnormal and/or infinite solutions which are typically encountered in previously published methods that use the rooftop primitive adjacency matrix to solve the critical rooftop vertices. In particular, first, an improved random sample consensus (RANSAC) algorithm is proposed to segment the rooftop primitives, i.e., the planar patches that constitute rooftops, of each building or group of connected buildings. The algorithm successfully maintains topological consistency among primitives and avoids under-and over-segmentation with high efficiency. Second, a novel Voronoi-based primitive boundary extraction algorithm under constraints of outer and inner building boundaries is introduced in order to extract each primitive boundary. In this algorithm, the adjacent segmented primitive relationships among the various primitives are preserved by a subgraph of the Voronoi diagram so that the reconstructed neighbor primitives are seamlessly connected. Third, in order to refine the boundary shapes of primitives with irregular geometry, various criteria for making the boundary adjustments more effective are proposed. In this way, more regular 3-D buildings can be produced. Finally, the primitive boundary simplification criteria are formally introduced to generate compact 3-D building models. By using the simplification criteria, nonadjacency between neighbor primitives, intersection between boundaries, and self-intersections are, to a great extent, avoided. Numerous experimental results obtained using multiple data sets, including data from the cities of Toronto and Enschede as well as from the Niagara area, have shown that the proposed methodology has excellent performance and it can produce watertight 3-D polyhedral building models.Index Terms-Airborne laser scanning (ALS), building boundary extraction, rooftop segmentation, three dimensional (3-D) building reconstruction, topological consistency, Voronoi-based diagrams.

A Review Of Point Clouds Segmentation and Classification Algorithms

“…A popular strategy in reverse engineering involves locally fitting primitives like planes, cylinders, cones using RANSAC based methods (Schnabel et al 2009). Chen et al (2014) has proposed a modified RANSAC segmentation algorithm that is less sensitive to noise, maintains topological consistency, and avoids over and under-segmentation of building primitives, through the localized sampling to segment the polyhedral rooftop primitives and then through the application of a region growing based triangulated irregular network (TIN) to separate the coplanar primitives. Several extensions are available within the Point Cloud Library:…”

Slicing Method for curved façade and window extraction from point clouds

“…Contemporaneous to that, Chen et al (2014) proposed an automatic and threshold-free evaluation system that offers an object-based technique for roof extraction. The enhanced algorithm performs well, even in the presence of noisy data and roofs covered in vegetation.…”

Towards the Automatic 3d Parametrization of Non-Planar Surfaces From Point Clouds in Hbim Applications

“…The automatic reconstruction of 3D models has been a challenging research topic and from the early beginnings numerous research papers based on different reconstruction methods have been published (Grün, 1995;Peternell, 2003;Pu, 2007;Pu, 2009;As Habib, 2010;Haala, 2010;Xiong, 2013;Cheng, 2014;Zolanvari, 2016;Yang, 2018). Even though a wide variety of modelling techniques ranging from nearly manual to almost automatic have been developed (Ochmann, 2016;López, 2018), automatic reconstruction still remains a major topic in which experts are putting their efforts on.…”