Innovative digital systems for the diagnosis of built heritage

The conservation and enhancement of historical architectural heritage play a crucial role in preserving the cultural memory and the social role of communities. Diagnosis and monitoring activities aimed at planning restoration and maintenance of the built heritage, which is often characterised by deterioration and structural instability, require time-consuming and costly manual labour. In recent years, the advent of advanced digital technologies has brought transformative changes to the field of architectural conservation. Techniques such as laser scanning, photogrammetry, Artificial Intelligence (AI), Internet of Things (IoT) systems and Historic Building Information Modeling (HBIM) environments offer effective solutions for documenting, analysing and monitoring the conditions and deterioration of historic buildings. Nevertheless, despite these technological advances, significant challenges remain, including the absence of standardised protocols, limited data interoperability, and the complexity of conducting interdisciplinary analyses. In this context, the research aims to establish an innovative methodological framework for the diagnosis and monitoring of the built heritage, with particular focus on reverse engineering techniques, Artificial Intelligence and IoT systems. The proposed approach aims to optimise the processes of data collection, analysis, and management of diagnostic information by leveraging reality-based three-dimensional models (point clouds and textured meshes), to conduct visual assessments of artefact’s state of conservation, including periodic monitoring over time. The methodology adopted involves data acquisition through photogrammetric surveys to generate high-resolution three-dimensional models, followed by the automatic classification of degradation pathologies by means of AI techniques, through the development ofmachine learning and deep learning algorithms designed to analyse image-based data derived from digital surveys (textures, UV maps, orthoimages). In this way, the threedimensional models are not merely a geometric reference for the 3D modelling phase, serving instead as intrinsic containers of both colorimetric and geometric information that also spill over into the two-dimensional data derived from them. In this way, the extraction of geometric and metric data allows for qualitative and quantitative analyses of the state of preservation in order to provide consistent and reliable diagnostic information, both in terms of the extent and the severity of degradation pathologies. Furthermore, integrating this information, along with sensor data from IoT systems, into interoperable and shareable HBIM-based environments allows for a complete and dynamic digital representation of the building. These representations support ongoing monitoring and evaluation phases of decay progression over time, enabling the planning of targeted and shared interventions. The research aims to establish a standardised and efficient methodological-operational workflow for the diagnosis of built heritage, promoting the replicability and scalability of the approach. Validation has been conducted through significant case studies about the historical cultural heritage of the Italian and Spanish territories. Based on the materials and structural characteristics, as well as the current conditions of the studied sites, different techniques of automatic classification of the pathologies have been validated and tested. The results have been integrated into semantically enriched digital environments, offering a solid basis for future technological and methodological advancements in the field of digitisation and architectural conservation.