Facade geometry generation from low-resolution aerial photographs for building energy modeling

Publication date: October 2017
Source:Building and Environment, Volume 123
Author(s): Jun Cao, Henning Metzmacher, James O'Donnell, Jérôme Frisch, Vladimir Bazjanac, Leif Kobbelt, Christoph van Treeck
In the Architecture, Engineering, Construction, Owners and Operators (AECOO) industry, geometry modeling of the Building Energy Modeling (BEM) project is often carried out by mechanical engineers and others who because are not versed in issues of building geometry definition and are not experts in the use of CAD software, often resulting in geometry definitions that do not adequately represent the true geometry of the modelled building and may result in failure or abandonment of the needed BEM and simulation project. In order to resolve the problem, this paper presents an automatic geometry modeling procedure of existing building facades, based on low-resolution aerial urban photographs, in order to recover their semantic structure as well as to synthesize geometric models for reuse in the BEM process. An iterative approach for building facade parsing is developed through the detection of facade structures. This approach is robust to image noise and under-sampling artifacts as they are common in aerial photographs. Based on this approach, the modeling system recovers high-level facade topologies and generates the geometry models of BEM for existing building facades. The geometry model generated is encoded into common BEM formats. Moreover, synthesis of realistic facade textures of the geometry model is achieved by using the City Geography Markup Language (CityGML). Finally, examples of energy simulation based on recovered BEM models are given illustrating the effectiveness of the proposed method and validating the process using real data.