Generalising 3D Buildings from LoD2 to LoD1

Generalising 3D Buildings from LoD2 to LoD1

Labetski, A. (TU Delft Urban Data Science)
Ledoux, H. (TU Delft Urban Data Science)
Stoter, J.E. (TU Delft Urban Data Science)

2017

The increasing popularity of 3D city models in navigation, urban planning, etc.,
necessitates application-specific and geometrically accurate and valid models. The concept of Levels of Detail (LoDs) indicate a model’s scale of adherence to its real-world counterpart. Highly detailed datasets often contain errors or require an exorbitant level of computing power. Given the high availability of LoD2 datasets, our research focuses on three considerations for generalising to LoD1: the vertical reference, extrusion vs. downtrusion and floor plan simplification. We present in this paper an initial methodology that produces geometrically accurate LoD1 models with a reduction of over 70% of the original file size.

Generalisation
3D
City Model
CityGML
Levels of Detail

http://resolver.tudelft.nl/uuid:5aaa4248-ef90-49ff-a238-f7b1dba5384e

University of Manchester

GISRUK 2017 Proceedings: 25th GIS Research UK Conference

GISRUK 2017: 25th GIS Research UK Conference, 2017-04-18 → 2017-04-21, Manchester, United Kingdom

Institutional Repository

conference paper

© 2017 A. Labetski, H. Ledoux, J.E. Stoter