Data Fusion Using Spot and Sar Images for Bridge and Urban Area Extraction

This paper is dedicated to bridge detection and urban area extraction from remotely sensed scenes. Taking into account the best potentialities, but also the limits of each sensor, our objective is to make an architecture that improves the performance and the reliability of scene analysis, compared to an architecture using a single sensor. For bridge detection, we first segment water in the SPOT image, to spa tially constrain the bridge research in the SAR image. This research is achieved using a correlation method. To detect an urban area, we first exploit the knowledge that it produces bright texture in SAR imagery, and a coarse mask is extracted using an adaptative thresholding in the SAR image. This mask is then used for classification as a training mask of urban area texture in SPOT image. We determine the non urban zone training set using a distance map of the urban training zone boundaries. Classification is performed with a multivariate Gaussian classifier.The results we obtained are very encouraging , especially if we consider the robustness of the bridge detection method.