ABSTRACT
Building Information Modeling is a powerful technology, but transferring information among applications is still limited by the diversity of their internal representation schema. The goal of real ‘Open BIM’ remains elusive due to the difficulty of the interoperability problem. For similar reasons, model-checking and functional simulations using building models is hampered by the need for careful tailoring of the content of model files exported for these purposes.
Semantic enrichment is a novel approach to this problem. It aims to apply expert system technology to interpret and enrich the semantic content of models so that they can be re-used for multiple purposes with minimal rework. The technique will have application in a wide variety of situations. Among those being developed in current research are precast concrete detailing, cost estimation, compilation of as-built models of bridges for bridge surveys, and acquisition of building data for facility maintenance.
In the context of the SeeBridge project, funded within the EU Infravation Program, the team is using a semantic enrichment prototype named SeeBIM 2.0 to aid in compiling building models of highway bridges from point cloud data that can be used for survey and recording of damage to the bridges. This large scale experimental application has yielded important insights into the ways in which rule sets can be compiled rigorously, to ensure that they can uniquely identify bridge components and the semantic relationships between them. The method, which uses feature vectors and feature relationship matrices, also suggests that the potential exists for an alternative approach using tools developed for computer vision that include machine learning.
Prof. Sacks’ talk will explore the need for semantic enrichment, its technology aspects, the ways in which it can contribute to providing interoperability, and the promise of a more advanced approach.
