Integration of HBIM and Machine Learning Processes for Cultural Heritage Maintenance

The technological development in the AEC sector is also radically influencing the area of cultural heritage. The processes of digitisation of the historical architecture are giving great pulse to projects aimed at physical and cultural accessibility, conservation, tourist attractiveness, dissemination and scientific research with positive effects in terms of enhancement and visibility. 3D survey, H-BIM and Scan-to-BIM processes, Machine Learning and Artificial Intelligence applications are contributing to a methodological and design approach revolution referred to CH. This paper shows a workflow related to the analysis of the degradation of ancient surfaces and its digital transposition on parametric models through the semi-automatic segmentation of HR images. The goal is to provide a useful workflow for populating ultra-specialized information in BIM environment for maintenance and restoration activities. The process is based on the following steps: laser scanning and photogrammetric survey of the object of analysis; Scan-to-BIM modeling; high resolution orthoimage editing of degraded surfaces; training of ML algorithms for automatic recognition of some degradation phenomena; production of info-graphic output of significant degradation; redaction of maintenance plan with related specialist attachments; integration of the results in BIM environment. The case study is a part of the western fortifications of the Aragonese Castle of Ischia; these protect Piazza d’Armi, are overlooking the sea and incessantly subjected to the corrosive actions of the wind, salt and rain. The object has been identified for its intrinsic inaccessibility and the consequent difficulties to intervene with traditional survey and analysis of degradation approaches.