Classification and Recognition Approaches for the BIM Modeling of Architectural Elements

Parametric and algorithmic modeling stands as an increasingly diffused key tool for virtualization and automation processes, also aimed at the cultural heritage's management and fruition. Indeed, while Building Information Modeling (BIM) notoriously represents the most suitable solution for the realization of digital models that serve as digital twins of new and existing buildings, there is an evident need to implement automatization processes of operation flows to populate BIM models. This refers to both their information content and geometric component: for the virtualization of existing buildings, the latter is tackled through the automation of the recognition processes for unstructured data from digital surveys, and the related generation of solid instances in BIM platforms, opening up to experimentations with machine learning techniques. Hence, in the context of this research theme, this contribution is focused on the construction of rules to characterize the data associated with the element, with a focus on the recognition and digitalization of architectural elements, outlining our recent research path on new procedural aspects in Cloud to BIM for the recognition and reconstruction of complex objects and forms from both images and point clouds. The paper will present some case studies, detailing specific innovative approaches with visual scripting and programming, highlighting the future research perspective to sample the potential of automatic classification through artificial intelligence algorithms.