Point Cloud Segmentation Based on the Uniclass Classification System with Random Forest Algorithm for Cultural Heritage Buildings in the UK

This paper presents an advanced hierarchical classification framework using the Random Forest (RF) algorithm to segment and classify large-scale point clouds of heritage buildings. By integrating the Uniclass classification system into a multi-resolution workflow, the research addresses key challenges in point cloud classification, including class imbalance, computational constraints, and semantic overlap at coarse resolutions. It adopts an experimental research design using the heritage case study from Royal Greenwich Museum in the UK. The findings demonstrate that industry classification systems and data taxonomies can be aligned with machine learning workflows. This study contributes to Heritage-Building Information Modelling (HBIM) by proposing optimised hierarchical structures and scalable machine learning techniques. The research concludes with recommendations for future research, based on the performance of the Random Forest technique, particularly in further developing AI applications within HBIM.