Filling holes in LoD2 building models

TL;DR

This work tackles automated hole filling in LoD2 city-building meshes plagued by geometric inaccuracies and topological defects. It introduces a three-phase pipeline—pre-processing to fix topology and identify pseudo-holes, hole-detection to extract complete border rings, and remeshing to reconstruct surfaces—without strict input requirements and with preservation of the original geometry, including explicit discrimination between true holes and pseudo-holes. Key contributions include a robust topological repair workflow using a half-edge framework, projection-based hole detection with thresholds $D<\epsilon_d$ and $A>\epsilon_t$ (where $\epsilon_d=0.1$ m and $\epsilon_t=0.01$), and a constrained Delaunay remeshing step on projected hole rings. Compared with established methods, the approach yields more complete and geometrically uniform hole fills, albeit with longer runtimes, and identifies several challenging scenarios that guide future work toward watertight, 2-manifold LoD2 models with correctly oriented normals.

Abstract

This paper presents a new algorithm for filling holes in Level of Detail 2 (LoD2) building mesh models, addressing the challenges posed by geometric inaccuracies and topological errors. Unlike traditional methods that often alter the original geometric structure or impose stringent input requirements, our approach preserves the integrity of the original model while effectively managing a range of topological errors. The algorithm operates in three distinct phases: (1) pre-processing, which addresses topological errors and identifies pseudo-holes; (2) detecting and extracting complete border rings of holes; and (3) remeshing, aimed at reconstructing the complete geometric surface. Our method demonstrates superior performance compared to related work in filling holes in building mesh models, achieving both uniform local geometry around the holes and structural completeness. Comparative experiments with established methods demonstrate our algorithm's effectiveness in delivering more complete and geometrically consistent hole-filling results, albeit with a slight trade-off in efficiency. The paper also identifies challenges in handling certain complex scenarios and outlines future directions for research, including the pursuit of a comprehensive repair goal for LoD2 models to achieve watertight 2-manifold models with correctly oriented normals. Our source code is available at https://github.com/tudelft3d/Automatic-Repair-of-LoD2-Building-Models.git

Figures (10)

• Figure 1: (a) displays a Lod2 building model reconstructed automatically, while (b) presents a manually reconstructed LoD2 building model.
• Figure 2: The LoD2 building mesh exhibits several errors, including non-manifold vertices highlighted in green, faces with incorrect normals indicated by the back faces of the mesh depicted in light blue, and boundaries of holes marked in red.
• Figure 3: Pipeline for hole filling in building mesh models.
• Figure 4: A 2D illustration depicts true and pseudo-holes in a 3D mesh. The left figure shows a true hole, while the middle and right display pseudo-holes. Red edges outline the hole's boundaries, and the light blue area indicates the real gap. Vertices in green highlight duplicate presence.
• Figure 5: The left figure illustrates the input mesh with self-intersecting faces in yellow and the boundaries of true and pseudo-holes in red. The right figure displays the output mesh, highlighting duplicated vertices in green, duplicated edges in yellow, and maintaining the remaining hole boundaries in red.