Geospatial information is indispensable for various types of spatially informed analysis and decision-making, such as traffic analyses, and natural resource management. In addition, geospatial information is one of the most powerful information integrators to bridge diverse sources of information. Such natures of geospatial information entail the need of geospatial data integration and geospatial knowledge outreach.
In most of real-world applications, geospatial data from one single source can hardly suffice. Therefore, integrating multi-source geospatial data is a predominant need for a variety of applications. However, today’s solutions for geospatial data integration in spatial data infrastructures (SDIs) are inadequate, and the data are often stored in the so-called “data silos”, i.e. datasets are stored mostly isolated from each other.
Semantic Web technologies provide a promising way for geospatial data integration on the Web. In this thesis, Semantic Web technologies are utilised to integrate multi-source geospatial data or integrated geospatial data with data from other domains. Paper I leveraged Linked Data and ontologies to realise a relative positioning approach, which positions thematic data based on their relations with background data. The relatively positioned thematic data can be automatically synchronised in terms of their geometric representations in all scales to avoid substantial discrepancy. Paper II integrated distributed multi-scale building data and a heritage building dataset to accomplish a heritage building map with both fine geometries and thematic information of heritage building. Paper III identified that only using ontologies is inadequate for integrating geospatial data and data from other domains, where complex and subtle semantic relations often arise. Then, a knowledge-based framework coupling ontologies and semantic constraints is developed to tackle the complex semantic relations raised by multiple representations of geospatial data.
Besides data integration, there is another prominent need for utilising geospatial data for various applications, that is, the outreach of geospatial knowledge. Visualisation, as one of the most predominant ways of utilising geospatial data, pertains to a wide range of cartographic knowledge. Therefore, it is desirable to formalise the knowledge of geospatial data visualisation (geovisualisation) to facilitate its interpretation, transfer, and reuse.
In this context, Semantic Web technologies offer a framework to formalise and share geovisualisation knowledge, thanks to their knowledge representation capacity. To this end, Paper II and III formalised the knowledge of geovisualisation in knowledge bases with ontologies and semantic rules. Such knowledge bases are evaluated in two real-world applications, i.e. heritage building mapping, and urban bicycling suitability mapping. The knowledge bases for geovisualisation can be used as a visualisation enablement layer for geospatial Linked Data.
In addition, Paper IV performed a study for the technical environment of the support for geospatial Semantic Web (Linked Data). It assessed and benchmarked several well-known and mainstream Linked Data stores, mainly in terms of their spatial query capacities and standard compliance. The results demonstrated that the support for geospatial Linked Data and queries has becoming increasingly mature. Nevertheless, query correctness remained a challenge for cross-database interoperability.
In conclusion, this thesis provides insights into the potentials of Semantic Web technologies for geospatial data integration and knowledge outreach (sharing). The insights could benefit the development of the next generation of SDIs, in which Semantic Web and Linked Data will expectedly play a role. From the perspective of Semantic Web research, the thesis contributes to the modelling and representation of geospatial data and knowledge on the Semantic Web.
Place: Pangea Auditorium, Geocentrum II, Sölvegatan 12, Lund
Name: Bernard, Lars
Affiliation: Technical University of Dresden, Dresden, Germany
- Earth and Related Environmental Sciences
- Geospatial data integration
- data visualisation
- Semantic Web
- Linked Data