SAR simulation of natural landscapes

This paper considers simulation of the SAR raw signal of natural landscapes. Within this framework, the format of the input data is critical. In fact, they are usually described by digital input maps which are too scarcely sampled for the application we are concerned. In the authors' case, they deal with digital elevation models, i.e., with the elevation input data. In order to interpolate such a data they employ the classical cubic spline interpolation and a fractal method, the latter based on the random midpoint displacement technique. Corresponding SAR images are generated and presented. Objective norms are further employed in order to compare simulated results with actual ones.<<ETX>>