From sketch BIM to design BIM: An element identification approach using Industry Foundation Classes and object recognition

Building Information Modelling (BIM) is a promising technology for building lifecycle management. Currently, BIM has been extensively used in both the sketch design (sketch) and detailed design (design) stages. Due to different purposes and different BIM tools, the reusability of sketch BIM models in the design stage is still challenging. To address this issue, this study proposes a building element identification scheme using a segmentation-aggregation strategy, termed EI-SA. Since Industry Foundation Classes (IFC) is the international standard of BIM, the EI-SA is built according to the IFC specification. Firstly, the element identification Problem is formally defined, followed by the overall framework of the EI-SA. Then, a presentation-level BIM segmentation algorithm is developed to divide a sketch BIM model into slices of geometric representations. Thirdly, a representation aggregation scheme is proposed to solve the element identification problem. The representation aggregation scheme employs a triangle-triangle intersection to detect the geometric connection and an object recognition algorithm to identify elements. Finally, experiments were conducted on several sketch BIM models. The experiment results showed that the EI-SA identified 82.22% elements on average, which verified the effectiveness of the EI-SA. To the best of our knowledge, this study is the first to solve the building element identification problem according to IFC. The proposed EI-SA will bridge the gap between sketch BIM models and design BIM models, smooth the reusability of data from the sketch stage to downstream stages, and improve the design efficiency by avoiding the re-designing of building elements.

Automated summary

This study proposes a building element identification scheme EI-SA using a segmentation-aggregation strategy, which effectively addresses the reusability challenge of sketch BIM models in the design stage. Experimental results verify the effectiveness of EI-SA in identifying building elements.