Semantic mediation of observation datasets through Sensor Observation Services

The large volume of environmental observation data that is currently being produced by public and private initiatives is highly heterogeneous both in data formats and semantics. A major challenge is to make all these data accessible in a uniform and integrated manner. Sensor Web Enablement specifications of the Open Geospatial Consortium provide standard data encodings and web service interfaces. However, semantic data mediation is still an issue to be solved. This paper describes a first effort for the semantic mediation between heterogeneous environmental observation datasets through Open Geospatial Consortium standard Sensor Observation Services. The solution enables application domain experts to provide semantic data integration knowledge within the scope of two well known top-level ontologies, namely W3C Semantic Sensor Network and NASA Semantic Web for Earth and Environmental Terminology. Such knowledge is combined with data source knowledge during the evaluation of global Sensor Observation Service GetObservation requests. Finally, the solution follows a Local As View data integration approach, thus new data sources may be incorporated without having to change the available data integration knowledge.