A model-driven methodology for integrating heterogeneous 3D geospatial urban entities

Abstract. Data on geospatial entities is increasingly available from various fields, such as GIS, CIM (City Information Modeling), and BIM (Building Information Modeling). They are described using different data models, such as IFC and CityGML, and are composed of both semantic and geometric data. Integrating this heterogeneous data to create a unified view of geospatial entities requires the use of technologies and methods adapted to each type of data. In this work, we propose a data integration methodology that leverages geospatial and semantic web technologies to provide data views suited to facilitating spatial and semantic navigation of these data. Concerning semantic data, we propose a model-driven data transformation process for limiting data loss during integration and reducing semantic heterogeneity between data from different urban information domains. For geometric data, we propose data transformations towards data formats for efficiently sharing, storing and visualizing 2D and 3D data. An open-source application is presented to illustrate this methodology.