Image Analysis Platform for Data Management in the Meteorological Domain

This paper proposes an architecture to provide semantic media information to the current existing meteorological models and prediction techniques. Satellite images have been used by meteorologists during the last 50 years, but we present a new method to take advantage of local images taken from the Earth. Networked terrestrial weather stations can offer valuable image information, both of local and wide areas adding details that cannot be captured by satellites. Based on the results of two projects carried out together with the Basque Meteorology Agency (Euskalmet), we propose a method to port from image data to semantic meteorological information and an architecture to integrate the existing weather data and knowledge structures with multimedia semantics. The validation of the analysis system has been carried out using sky images taken in visual spectrum and the results have demonstrated the great potential of such platforms that could be extended to other data sources in order to apply multimedia semantic technologies in application fields like meteorology.