Cross-Disciplinary Semantic Building Fingerprints ‚ AI Knowledge Graphs to Store Topological Building Information Derived From Semantic Building Models (BIM) to Apply Methods of Artificial Intelligence (AI)

Abstract

The advancing digitalization in the building sector with the possibility to store and retrieve large amounts of data has the potential to digitally support planners with extensive design and construction information. Large amounts of semi-structured three-dimensional geometric data of buildings are usually available today, but the topological relationships are rarely explicitly described and thus not directly usable with computational methods of AI. To this end, we propose methods for indexing spatial configurations inspired by the similarity analysis of incomplete human fingerprints, since the early design stage of architectural design is characterized by incomplete information. For this, the topology of spatial configurations is extracted from Building Information Modelling (BIM) data and represented as graphs. In this paper, Semantic Building Fingerprints (SBFs) and Semantic Urban Fingerprints (SUFs), as well as use cases for AI methods are described.