Automatic Generation of Geospatial Metadata for Web Resources

Web resources that are not part of any Spatial Data Infrastructure can be an important source of information. However, the incorporation of Web resources within a Spatial Data Infrastructure requires a significant effort to create metadata. This work presents an extensible architecture for an automatic characterisation of Web resources and a strategy for assignation of their geographic scope. The implemented prototype generates automatically geospatial metadata for Web pages. The metadata model conforms to the Common Element Set, a set of core properties, which is encouraged by the OGC Catalogue Service Specification to permit the minimal implementation of a catalogue service independent of an application profile. The performed experiments consisted in the creation of metadata for Web pages of providers of Geospatial Web resources. The Web pages have been gathered by a Web crawler focused on OGC Web Services. The manual revision of the results has shown that the coverage estimation method applied produces acceptable results for more than 80% of tested Web resources.