Robust and fast reconstruction of complex roofs with active sampling from 3D point clouds

Abstract

This article proposes a novel method for the 3D reconstruction of LoD2 buildings from LiDAR data. We propose an active sampling strategy which applies a cascade of filters focusing on promising samples at an early stage, thus avoiding the pitfalls of RANSAC‐based approaches. Filters are based on prior knowledge represented by (nonparametric) density distributions. In our approach samples are pairs of surflets—3D points together with normal vectors derived from a plane approximation of their neighborhood. Surflet pairs provide parameters for model candidates such as azimuth, inclination and ridge height, as well as parameters estimating internal precision and consistency. This provides a ranking of roof model candidates and leads to a small number of promising hypotheses. Building footprints are derived in a preprocessing step using machine learning methods, in particular support vector machines.

Reconstruction of Complex Roof Semantic Structures from 3D Point Clouds Using Local Convexity and Consistency

“…Poullis [33] developed a complete framework to automatically reconstruct urban building models from point clouds by combining a hierarchical statistical analysis of the data geometric properties and a fast energy minimization process for the boundary extraction and refinement. To generate more detailed roof models, Dehbi et al [34] propose a novel method for roof reconstruction using active sampling, and it is limited to only dormer types. The main advantage is that it can reconstruct a polyhedral building with complex shapes, while the main drawback is the sensitivity to the incompleteness of the point cloud caused by occlusions, shadows, etc.…”

“…The model-driven approaches can robustly reconstruct building models with simple roof styles by utilizing prior knowledge like parallel and symmetry, generating watertight building models. However, roof primitives or structures in the true world reveal a huge diversity, thus, it will fail when a searched roof cannot be described by any of the predefined primitive [34,48,49].…”

“…The hybrid-driven 3D reconstruction approaches [10,34,48,50], combining the conventional data-and model-driven strategies, aim to recognize building roof structures or search the best-fitted roof primitives from a predefined roof library. These hybrid-based algorithms can benefit from the roof topology graphs (RTG), which is a graph reflecting neighboring relations between extracted roof planes.…”

Automatic Building Footprint Extraction From 3d Laserscans

“…Automatic processing of 3D point clouds of buildings has been an active field of research for years. This often involves the automatic identification and reconstruction of specific building elements, such as windows and doors (Pu and Vosselman, 2007), roofs (Dehbi et al, 2021), or extraction of the building's floorplan (Gankhuyag and Han, 2020). Since our approach aims at the automatic generation of the building footprints, we first highlight recent work in this area.…”

Intuitive and efficient roof modeling for reconstruction and synthesis

“…Lin et al [2013] propose to find a combination of planar primitives that best explains the input Li-DAR data for roof reconstruction. Dehbi et al [2021] propose an active sampling strategy for RANSAC to fit plane approximations to input point clouds. Nan and Wonka [2017] and Kelly et al [2017] employ integer programming to reconstruct coarse planar building models.…”