Transversal Knowledge in Teaching Architecture: Applications of 3D Point Cloud Segmentation Methods for the Comprehension of Cultural Heritage

Teaching innovation in architecture and building engineering aims to go beyond traditional methods by incorporating novel approaches, preparing students for the dynamic industry shaped by technology, sustainability, and design advancements. This innovation relies on the integration of technologies, interdisciplinary collaboration, sustainability, experiential learning, and critical thinking. In the field of architecture and engineering, the use of digital information systems, including photogrammetry and laser scanning, in documenting and analyzing cultural heritage, is growing. Although these techniques rapidly collect point clouds, transitioning them into 3D representations is a manual and tedious process. The present work recounts the experience gained during the realization of the 3D-HUB project of the University Polytechnic of Bari in collaboration with the engineering company St. Arch. Starting from a dialogue on the gaps of the Scan-to-BIM process within the historical heritage, two methods of segmentation of point clouds were examined, one traditional, through the use of the RANSAC algorithm, and the other innovative, based on machine learning. The outcome of this trial on the case study of the former orphanage Casa Santa in the historic center of Monopoli, Italy, has emerged considerations on educational innovation in future courses provided at the University Polytechnic of Bari.