Conference Agenda

Trasimeno Smart Land is a research initiative focused on the integration of 3D geospatial data to support sustainable territorial governance through a digital infrastructure derived from the built environment’s digitalisation. Applied to a vast territory in central Italy, administered by a union of eight small municipalities, the project represents an evolution of the Smart City paradigm within the broader framework of Smart Land. The district, characterised by extensive natural areas, rich cultural heritage, low-density settlements and a fragmented built environment, requires the adoption of innovative tools and approaches for its management due to the inherent fragmentation of the area. This research illustrates how limited urbanized areas (Cialdea, 2018) can be effectively mapped and managed for sustainable territorial planning (Gerli et al., 2024; Mohan et al., 2025), through advanced data governance and digital representation. This initiative led to the creation of PITER, “Integrated Territorial Platform”, ideated by the municipality union in collaboration with the University of Perugia and developed by three local companies. The digital system, integrating geospatial, environmental, urban and demographic data, enhances local administrative tools and – through iterative, municipality-driven validation – emerges as a replicable model that concretely operationalises the Smart Land concept (Azadi et al., 2023).

The project has led to the development of a Digital Twin, that covers the entire 780 square kilometres territory of the municipal union and is implemented through an interactive platform, which surpasses static datasets by incorporating real-time dynamic data via sensors and update systems, thereby significantly improving its analytical capabilities. The system is structured around three interconnected core components (fig. 1), designed to enhance territorial management and support the visualization and administration of the digitalized environment. These connections – encompassing territorial insight, citizen reports, points of interest, technical data, street asset surveys and cadastral mapping data – are structured across 24 thematic and functional layers, together forming an information ecosystem that facilitates intelligent and efficient territorial governance. The innovative services provided by the Digital Twin are the outcomes of a new approach directed towards digital transition. PITER, an interrogable repository and infrastructure (fig. 2), acquires processes and provides geospatial data in a fully usable format. Designed to ensure high interoperability between technological components, it supports informed, data-driven decision-making. The first component of PITER is WiseTown Citiverse, integrated with ArcGIS’s GIS engine, which hosts a dynamic digital model of the lake district, designed to support complex spatial analyses and territorial planning scenarios. The model constitutes a fully operational Digital Twin within PITER’s interface, serving as a comprehensive repository of territorial data that consolidates diverse datasets from multiple sources, including geospatial data on the built environment, urban tree mapping and key infrastructure like healthcare facilities and other points of interest. It also integrates municipal planning documents and technical environmental maps, such as the Regional Technical Map provided by the regional administration, as well as landscape protection zones for wooded areas and watershed protection areas outlined in Provincial Territorial Coordination Plans. A specific dataset, resulted from survey campaigns covering approximately 380 kilometres of urban roads, including nearly 130 000 collected images and 16 452 237 749 point cloud data points, has been incorporated into the system. Through in-depth analysis, this data has enabled the precise mapping of drainage inlets (approximately 20 per kilometre), trees (around 27 per kilometre), vertical signage and 18 000 residential building numbers, further enriching the platform’s capabilities. Key features include a construction site monitoring system that provides updates on active public and private projects and the World Traffic Service, which tracks vehicular flow, car accidents and congestion. Additional thematic layers support broader land-use governance, such as maps of major tourist attractions and digitalized, georeferenced cadastral records. Many of these datasets are continuously updated and available for detailed consultation. The platform also incorporates the WiseTown Issue Manager, a participatory tool that fosters citizen engagement in territorial governance, allowing residents to report urban management, infrastructure and public service issues while tracking intervention progress. Reported cases are seamlessly integrated into the Digital Twin, enabling local administrations to efficiently visualize and address concerns within the digitalized environment. The third key component of PITER is a suite of Thematic Dashboards that aggregate and analyse essential territorial data. These dashboards present welfare indicators and urban greenery management data from ISTAT (Italian National Institute of Statistics) census and environmental monitoring data collected from IoT sensors in weather stations across the region, granting statistics evaluation. Beyond these core functionalities, the project includes digital tools designed to promote and manage the territory. The Grand Tour Trasimeno website showcases the region’s cultural and environmental heritage and is directly linked to a dedicated Digital Twin layer featuring the same points of interest. Additionally, the Geolander.it portal provides access to georeferenced data and point clouds from territorial surveys, supporting analysis and planning efforts. This interconnected ecosystem strengthens the platform’s ability to meet diverse stakeholder needs, bridging technological innovation with sustainable development strategies.

Trasimeno Smart Land represents a pioneering initiative with significant potential to optimise public administration processes. Designed as a dynamic and adaptable system, it continuously evolves through the integration of new data, many of which are real-time, such as traffic data and construction site updates, and modifiable by administrators via ArcGIS, such as municipal territorial maps that are regularly revised and expanded as new ones are digitized. This flexible data structure allows seamless integration with other services within the platform, creating a synergistic system for territorial planning, management and monitoring. Ultimately, the PITER platform not only establishes a solid foundation for future innovations, such as the integration of artificial intelligence and virtual reality, but also redefines the way to approach territorial governance. By transforming data into an accessible and actionable resource, it empowers more informed decision-making and streamlines administrative processes. The research bridges the gap between Smart City practices and extended fragmented areas, tackling the greater complexities of data and territory management beyond urban confines through the development of PITER as a pioneering model of Smart Land.

Embedding spatial intelligence into vision‐language models (VLMs) has offered a promising avenue to improve geospatial decision‐making in complex urban environments. In this work, we propose a novel framework that augments the architecture of Contrastive Language-Image Pretraining (CLIP) with the techniques of spatial-context aware prompt engineering and spatially explicit contrastive learning. By leveraging a diverse set of geospatial imagery (e.g., street view, satellite, and map tile images), paired with contextual geospatial text generated and curated via GPT-4, our approach constructs robust multimodal representations that capture visual, textual, and spatial insights. The proposed model, termed GeospatialCLIP, is specifically evaluated for urban mixed land use detection, a critical task for sustainable urban planning and smart city development. Results demonstrate that GeospatialCLIP consistently outperforms traditional vision-based few-shot models (e.g., ResNet-152, Vision Transformers) and exhibits competitive performance with state-of-the-art models such as GPT-4. Notably, the incorporation of spatial prompts, especially those providing city-specific cues, significantly boosts detection accuracy. Our findings highlight the pivotal role of spatial intelligence in refining VLM performance and provide novel insights into the integration of geospatial reasoning within multimodal learning. Overall, this work establishes a foundation for future spatially explicit AI development and applications, paving the way for more comprehensive and interpretable models in urban analytics and beyond.

In subsurface planning of urban areas, the use, management, and exchange of urban and geological data serve as a fundamental basis for collaboration among various specialist in the fields of urban and geological 3D modeling. Therefore, the development of a data model schema tailored to urban and geological subsurface models is an important foundation for data management. However, various information’s about buildings, infrastructures, and geological structures, that cannot be integrated in common data model structures requires an expansion and combination of the existing data model structures for geological and urban 3D models. The existing data model schemas can already manage various types of subsurface data. Building on this, this paper aims to advance the development of a data model by combining existing schemas to meet the requirements for data management in 3D voxel modeling of the subsurface. The conception of such a data model is implemented for a case study in the City of Stuttgart, based on the GML application schemas Geoscience Markup Language (GeoSciML) and City Geography Markup Language (CityGML). Our data model is applied using urban and geological model elements for the Stuttgart case study. In this context, the construction and integration of the model elements within a framework that comprises modeling, data management, and visualization tools were examined.