3D Point Cloud Annotation through an ontology model: the case-study of San Giovanni a Porta Latina, Rome

Abstract. Ontologies play a crucial role in the semantic enrichment of 3D models, enabling structured knowledge representation and interoperability in the digital documentation of built heritage. This study presents a multi-layered semantic segmentation workflow for point clouds that integrates spatial classification with formal ontological modelling. The proposed method is applied to the Church of San Giovanni a Porta Latina in Rome, where the point cloud was manually segmented using the open-source software CloudCompare according to five thematic layers: constructive element, construction technique, material, chronological reference, and decay. Semantic labels were assigned through the Cloud Layers plugin and embedded into the ASCII-exported point cloud, forming the basis for conversion into RDF and integration into a semantic web framework. This structure supports querying via SPARQL, allowing for the extraction and analysis of complex relationships within the dataset. By working directly on the point cloud his approach surpasses the limitations of traditional 2D documentation, offering a more holistic understanding of architectural complexity. The proposed methodology provides a scalable and interoperable approach to cultural heritage documentation, enabling spatial reasoning and data-driven decision-making directly on digital twins of historic structures.