Processi cognitivi e infrastrutture digitali per la conoscenza e la narrazione del costruito storico. L'integrazione tra rilievo, modellazione e intelligenza artificiale.

This research critically and experimentally investigates contemporary processes of Reality-based representation oriented toward the knowledge, documentation, interpretation, and communication of the built heritage. The study develops through an articulated set of multisensor surveying practices, digital modelling procedures, lighting simulations and generative artificial intelligence techniques, taking complex historical architectures as its privileged field of inquiry. In such contexts, characterised by morphological discontinuities, temporal stratifications, material variations and irregular spatial configurations, both acquisition procedures and representational languages are challenged. The convergence between high-accuracy technologies and rapid, low-cost tools enables the construction of coherent, transparent, and traceable knowledge processes capable of returning phenomena that exceed the sole metric dimension. The research is situated within the Scientific Disciplinary Sector CEAR-10/A – Drawing, a field concerned with the generation, construction, and analysis of drawings, images and models as scalar representations of existing or designed reality. Within this disciplinary framework, the investigation operates at the intersection of the sector’s epistemic axes: the geometric–descriptive axis, devoted to the rigorous documentation of form and its relations; the informational–computational axis, focused on data structuring and on the interoperability between models and systems; the graphic–narrative axis, concerned with exploring the communicative, symbolic and interpretative capacities of representations. In developing these lines of inquiry, the dissertation also engages with themes belonging to the SSD IIND-07/B – Building Physics, particularly with regard to the modelling of luminous behaviours, physically reliable simulations, and the analysis of conditions related to natural and artificial lighting. This intersection makes it possible to extend the documentary apparatus into the domain of perceptual phenomena, translating the relationship between light, matter and space into forms of visualisation, without reducing such simulations to mere exercises in graphic hyper-realisation. The objectives of the research unfold along three main directions: assessing the effectiveness and limits of multisensory surveying techniques in representing the complexity of historic architectures, taking into account differences in scale, morphological articulation and material behaviour; critically analysing the transition from the digital model to architectural drawing, showing how automatic extraction procedures, at present, cannot replace the selective, normative and communicative choices inherent to disciplinary drawing and to its interpretative function; exploring forms of heritage narration and communication capable of integrating documentary rigour, cultural awareness and evocative capacity through plural visual languages, including those enabled by generative artificial intelligence. Methodologically, the work adopts an integrated approach based on three complementary perspectives: a historical–critical reconstruction of the genealogy of the model within architectural representation, aimed at clarifying continuities, ruptures and transformations in the paradigms of reality depiction; an experimental and comparative assessment of multisensory surveying strategies applied to different real contexts; and a comprehensive testing of the procedures through the case study of the monumental complex of Santa Maria della Vita in Naples. The research is grounded in the awareness that representation is never a passive process nor a mere technical translation. It constitutes a cognitive act that entails choices, interpretations and modes of constructing meaning, through which not only what is surveyed is defined, but also the way in which heritage is rendered legible and communicable. This awareness has guided the analysis of the plurality of methodologies employed, showing how each acquisition technique and each modelling paradigm incorporates distinct structures of meaning, operational constraints and interpretative possibilities. Within this framework, metric and informational modelling plays a central, though not exhaustive, role. The investigation assumes that the Reality-based model does not represent a provisional phase to be simplified or translated, but constitutes a cognitive device in its own right, capable of preserving geometric relations, spatial configurations and material qualities that would be inevitably reduced through immediate conversion into abstract models. Consequently, the transition toward information-rich environments — including the Mesh-to-HBIM process — is examined also in its cultural and interpretative dimensions, interrogating both the selective responsibility of the modeller and the role of automation. This structure gives rise to the methodological questions examined in the research: what role does interpretative selection play in defining representational media? How can the richness of the survey be translated into informational structures without diminishing its complexity? To what extent can automation coexist with authorial responsibility and with the cultural awareness inherent to drawing? Further issues emerge from the passage from the three-dimensional model to architectural drawing: automatic extraction procedures, although technically advanced, do not in themselves guarantee adherence to the graphic codes and disciplinary conventions of bidimensional representation. These elaborations still require interpretative choices, thresholds, conventions and capacities for synthesis that point to the indispensable role of drawing as a critical tool. In parallel, the research explores artificial intelligence in two complementary directions: the use of three-dimensional neural models (NeRF) to explore volumetric and perceptive continuities; the employment of text-to-image algorithms to render aspects of heritage that are not strictly metric, such as atmospheres, rituality, symbolic stratifications or the evocative qualities of space. These tools are not regarded as alternatives to disciplinary drawing practices, but as critical objects to be understood, assessed and consciously integrated into representational processes, ensuring control, legibility and documentary responsibility. In this manner, the case study of the complex of Santa Maria della Vita becomes a critical laboratory in which to observe continuities and ruptures between surveying, modelling and representational practices, and in which to measure the distance — or coherence — between what the survey records, what modelling formalises and what representation communicates. It is the locus where the entire theoretical and methodological structure of the research is tested through direct engagement with the complexity of the built environment. Multisensory surveying, informational modelling, lighting simulation and artificial intelligence experimentation are employed not as isolated procedures, but as interdependent components of a single knowledge system. Ultimately, the research proposes a flexible and critically grounded methodological structure for the construction of digital heritage ecosystems capable of combining documentary accuracy, critical awareness and narrative openness.