VOLUMINATOR 2.0 – SPEEDING UP THE APPROXIMATION OF THE VOLUME OF DEFECTIVE 3D BUILDING MODELS
Semantic 3D city models are increasingly used as a data source in planning and analyzing processes of cities. They represent a virtual copy of the reality and are a common information base and source of information for examining urban questions. A significant advantage of virtual city models is that...
| Main Authors: | , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Copernicus Publications
2016-06-01
|
| Series: | ISPRS Annals of the Photogrammetry, Remote Sensing and Spatial Information Sciences |
| Online Access: | http://www.isprs-ann-photogramm-remote-sens-spatial-inf-sci.net/III-2/29/2016/isprs-annals-III-2-29-2016.pdf |
| Summary: | Semantic 3D city models are increasingly used as a data source in planning and analyzing processes of cities. They represent a
virtual copy of the reality and are a common information base and source of information for examining urban questions. A significant
advantage of virtual city models is that important indicators such as the volume of buildings, topological relationships between objects
and other geometric as well as thematic information can be derived. Knowledge about the exact building volume is an essential base
for estimating the building energy demand. In order to determine the volume of buildings with conventional algorithms and tools,
the buildings may not contain any topological and geometrical errors. The reality, however, shows that city models very often contain
errors such as missing surfaces, duplicated faces and misclosures. To overcome these errors (Steuer et al., 2015) have presented a robust
method for approximating the volume of building models. For this purpose, a bounding box of the building is divided into a regular
grid of voxels and it is determined which voxels are inside the building. The regular arrangement of the voxels leads to a high number
of topological tests and prevents the application of this method using very high resolutions. In this paper we present an extension of the
algorithm using an octree approach limiting the subdivision of space to regions around surfaces of the building models and to regions
where, in the case of defective models, the topological tests are inconclusive. We show that the computation time can be significantly
reduced, while preserving the robustness against geometrical and topological errors. |
|---|---|
| ISSN: | 2194-9042 2194-9050 |