ABSTRACT
3D point cloud data describes our physical world spatially. Knowledge discovery processes including semantic segmentation and classification are a great way to complement this information by leveraging analytic or domain knowledge to extract semantics. Combining efficiently this information is an opening on intelligent environments and deep automation. This chapter provides a conceptual data model to structure 3D point data, semantics and topology proficiently. It aims at creating an interactive clone of the real world usable by cognitive decision systems. A multi-modal infrastructure integrating this data model is presented that includes knowledge extraction, knowledge integration and knowledge representation for automatic agent-based decision-making over enriched point cloud data. A knowledge base processing with ontologies is provided for extended interoperability.
