Water detection with high-resolution SAR image based on texture and imaging knowledge

In order to extract water accurately, we analyze the characteristics and the imaging mechanism of the country buildings in the synthetic aperture radar (SAR) images, and find that the water detection is interfered by the shadow of the buildings. Therefore, the water is accurately extracted from the high-resolution SAR image after getting rid of the building shadows by the texture method and the knowledge-based building representation. This paper indicates that it can more accurately extract information from the high-resolution SAR images with the traditional methods, combining with the knowledge-based methods. After extracting water we can analyze the correlative information farther.