Artificial Intelligence Enhances Building Information Modelling

Research from University Francisco de Vitoria in Madrid, Spain, addresses the challenge of how Artificial Intelligence (AI) can streamline the Scan-to-Building Information Modelling (Sc2BIM) process. The study concludes that AI can significantly enhance the Sc2BIM workflow by automating critical tasks such as point cloud segmentation and model generation. Financial support for this research came from the European Union (EU).

Key Takeaways:

• The Sc2BIM process is a traditional workflow that remains time-consuming and resource-intensive, with point cloud segmentation and model generation being critical tasks that require significant manual effort.
• AI techniques, particularly PointNet++ and Convolutional Neural Networks (CNNs), can significantly enhance the Sc2BIM workflow by automating these tasks and improving model accuracy.
• The study highlights the growing role of AI in streamlining the Sc2BIM pipeline, potentially reducing manual effort and improving model accuracy.
• The research identifies PointNet++ as the most frequently used model for 3D point cloud segmentation, while CNNs remain the dominant architecture overall.
• Hybrid and transformer-based models are gaining popularity in the field, and some studies demonstrate successful full 3D BIM reconstructions from raw scans.

Statistics:

• The study conducted a systematic review of 113 relevant papers from main scientific resources.
• PointNet++ is the most frequently used model for 3D point cloud segmentation.
• CNNs remain the dominant architecture overall, with a 75% adoption rate.
• Hybrid and transformer-based models have a 25% adoption rate.

Sources:

• A Systematic Review of Artificial Intelligence for Capturing Real-world Structures Into Building Information Modelling. Journal of Building Engineering, 2025;113.
• University Francisco de Vitoria, Madrid, Spain
• European Union (EU)
• Journal of Building Engineering, Elsevier's Radarweg 29, 1043 Nx Amsterdam, Netherlands