Host institution: Technion – Israel Institute of Technology
Project Title: Object relationships detection and semantic enrichment
Objectives: To devise and test two approaches to the semantic enrichment of BIM models: rule-based inferencing and machine learning with sparse data sets. To build prototypical software implementations of each approach, and possibly of a hybrid approach. To test the approaches using the prototypes to models of buildings and bridges as use cases, through applications such as the generic reconstruction of native BIM models, code compliance checking, and various simulation and analysis applications.
Expected Results: The ESR will implement and test a thorough proof of concept of the semantic enrichment approach to compiling BIM models from the geometry models produced from point cloud data; the proof of concept will cover three approaches (rule-based inferencing, machine learning and hybrid). The outcome is expected to include both a theoretical and a practical (empirical) evaluation and comparison of the efficacy of each approach. The software implementation will be tested on two construction domains (automated space use identification for code compliance checking, and model compilation for highway bridge survey). The report is expected to include detailed recommendations for industrial implementation. Results will be rigorously checked by automated analysis of resulting IFC instance files against formal Model View Definitions.
Enrolment in Doctoral degree: Technion – Israel Institute of Technology, PhD Civil Eng.
Advisor: Rafael Sacks