Learning filters in Gaussian process classification problems
This paper reports the results of UrbEx (urban expansion monitoring) project of the European Space Agency (ESA) carried out in collaboration with the Word Wildlife Fund (WWF) as end-user. The project, which has been carried out by the Italian company advanced computer system (ACS), was aimed at developing a user-oriented information service for monitoring urban expansion at national scale. Such a service was tailored to the requirements of the user, WWF, for supporting the conservation activities of this international nongovernmental organization. From a methodological viewpoint, the system was based on a novel data fusion approach to land-cover classification. In particular, both Landsat and 35-days interferometric ERS images where used to achieve an accurate detection of the urban tissue over the whole Italian coastal area at two different times, 1995 and 2000. The final results validated by WWF demonstrate the capabilities of EO technology to provide reliable, synoptic and continuous information on the urban expansion at national, regional and even global scale.