An interactive fuzzy fusion system applied to change detection in SAR images

The contribution of this paper concerns the detection of changes in multi-temporal satellite SAR (synthetic aperture radar) images by a fuzzy fusion of attributes extracted from the images with a-priori map-based information. The proposed approach is based on a linguistic description of the attributes and of the relations between them that are provided by geophysicists. The fusion system is cooperative thanks to a graphical user interface that allows one to visualize reference areas in the attribute space and to easily adjust some parameters of the attribute fusion (rules, membership functions). Some detection results of small deforested areas are presented.