Digital Heritage Documentation. Mapping Features Through Automatic, Critical-Interpretative Procedures

The contribution is focused on digital data segmentation to define materials, construction techniques, and state of conservation, linking these identified features to the H-BIM model. The main aim is to outline comparative thematic documentation of cultural heritage, up to the implementation of semantic web platforms for monitoring and management. The use of tools and procedures for digitising cultural heritage is becoming increasingly widespread. The latest 3D survey technologies are able to produce very accurate models in a very short time, with clear advantages in terms of speed and metric precision during onsite acquisition, while the processing of the acquired data can be very time-consuming and complex. These 3D models with a high information density allow a heritage understanding in its metric, morphological, structural, material and conservation features, but proper digital data management processes are needed to segment and discretize data through semantic classifications with a high critical-interpretive value. The purpose is to explore Artificial Intelligence (AI) processes in order to combine the irreplaceable cultural and interpretative skill with suitable tools useful to prioritising data in a hierarchical way according to different levels of knowledge through automatic procedures. The research is in its very beginning. Methodology is currently being set up, and some comparative data sets are being explored. A set of existing point cloud databases will be analysed, in order to point out classification criteria to outline specific features (construction techniques, surface features, state of conservation, morphologies, etc.). This will lead to a methodology that facilitates an increasingly structured organisation of interpretative data, also within shared platforms.