SAR Image Fusion in Multi Sensor Context for Small Urban Area Detection

The purpose of this work is to provide new fusion methods adapted to multi sensor SAR use for urban area segmentation. We explain first backscattering of radar signal from urban object and characterize variation due to sensors characteristics. Then, we propose two new data fusion methods suitable for radar images. The first one will use similar response to enhance quality of building backscattering and the second one will use two images to calibrate a statistical scheme based on the differences in backscattering from speckled area to urban area. We will show the results of our work applied on tropical areas.