Image-based Deep Learning Approaches for Point Cloud Classification for Heritage BIM Modelling

View Full Paper

Ask AI

Bookmark

Authors

Discussion

This analysis shows improved segmentation and semantic classification in heritage point clouds, suggesting new workflows for HBIM.

The hybrid pipeline produced spatially consistent 3D segments that reduced manual labor.
Results indicated that the Segment Anything Model generalises well with equirectangular perspectives, enhancing accuracy.
Experiments utilized high-resolution data from the Queen’s House, revealing improved semantic labelling with YOLO World.
Challenges like projection ambiguity indicate that further research is essential for practical HBIM adoption.

Gil et al. (2025) studied this question.

View Full Paper

Ask AI

Bookmark