As a domain extendable 3D city model the OGC standard CityGML has been increasingly employed as a dominant consensus over different application areas. An essential challenge encountered in many research and commercial activities in the field of 3D city modelling is to find a reliable approach for achieving high-efficient storage and management of data models according to CityGML with its Application Domain Extensions (ADE) in order to ensure interoperable data access across broad application domains. Based on graph transformation, this paper introduces a new approach along with an extensive database structure that allows for dynamically extending the spatially-enhanced relational databases for handling arbitrary CityGML ADEs by means of graph transformation systems. With this approach, relational database schemas with simplified and optimized table structures can be automatically generated from the XML application schemas of CityGML ADEs by performing a series of user-defined graph transformation rules which can describe complex mapping rules for transforming object-oriented data models to relational database models in a fully declarative way. The proposed approach has been successfully implemented and tested based on a number of different CityGML ADEs like Energy ADE, Dynamizer ADE, and UtilityNetwork ADE.
«