Data fusion and filtering via calculus of variations

We study the problem of urban areas detection from satellite images. In particular, we consider two types of satellite images: SAR (Synthetic Aperture Radar) images and optical images. We describe a simple algorithm for the detection of urban areas. We show that the performance of the detection algorithm can be improved using a fusion procedure of the SAR and optical images considered. The fusion algorithm presented in this paper is based on a simple use of ideas taken from calculus of variations and it makes possible to do together the filtering and the data fusion steps. Some numerical examples obtained processing real data are reported at the end of the paper. In the website http://web.unicam.it/matinf/fatone/wl several animations relative to these numerical examples can be seen. © 2006 International Federation for Information Processing.