Combining an Angle Criterion with Voxelization and the Flying Voxel Method in Reconstructing Building Models from LiDAR Data

Abstract

Traditional documentation capabilities of laser scanning technology can be further exploited for urban modelling through the transformation of resulting point clouds into solid models compatible for computational analysis. This paper introduces such a technique through the combination of an angle criterion and voxelization. As part of that, a k-nearest neighbor (kNN) searching algorithm is implemented using a predefined number of kNN points combined with a maximum radius of the neighborhood, something not previously implemented. From this sample points are categorized as boundary or interior points. Façade features are determined based on underlying vertical and horizontal grid voxels of the feature boundaries by a grid clustering technique. The complete building model involving all full voxels is generated by employing the Flying Voxel method in order to relabel voxels inside openings or outside the facade as empty voxels. Experimental results on 3 different buildings, using 4 distinct sampling densities show successful detection of all openings, reconstruction of all building façades, and automatic filling of all improper holes. The maximum nodal displacement divergence was 1.6% compared to manually generated meshes from measured drawings. This fully automated approach rivals processing times of other techniques with the distinct advantage of extracting more boundary points, especially in less dense data sets (< 175pts/m 2 ), which may enable its more rapid exploitation of aerial laser scanning data and ultimately preclude needing a priori knowledge.

Octree-based, automatic building façade generation from LiDAR data

“…The FD and FA approaches automatically reconstruct building, where the boundary points were extracted based on characteristic of triangles in triangulation mesh and an angle criterion, respectively. These algorithms are described elsewhere [6,7].…”

“…In conclusion, the FV algorithm can eliminate incorrect boundary points during reconstructing boundary lines. The efficacy of the FV approach and its compatibility with computational modelling were benchmarked against the commercial program Kubit [19] and two automatic approaches developed by the authors (the FacadeDelaunay (FD) algorithm [6] and the FacadeAngle (FA) [7]). The Kubit software allows users to use AutoCAD tools to create building models based on point clouds.…”

“…6,7]. As such, the FV algorithm is proposed both to improve accuracy of the resulting building models, while significantly reducing the computational cost.…”

“…doors or windows), if their characteristics satisfy those of common openings in building façades involving minimum lengths and widths equal or greater than 0.4m [6,7,12,21] and a height-to-length ratio between 0.25 and 5.0 [38,41] …”

“…Laser scanning is gaining popularity in reverse engineering [1][2][3] and visualization [4,5], yet exploiting the data for structural analysis remains a topic for development [6,7]. Developing automatic, robust, and efficient methods to reconstruct geometric building models for computational modelling has the potential to save money and time for projects needing cityscale modelling.…”