Non-Gaussian surface generation

An approach has been found to generate two-dimensional non-Gaussian surfaces with known height densities and surfaces with known height densities and surface correlation functions. Four types of surfaces with various combinations of Gaussian and non-Gaussian height densities and correlation functions are generated. Statistical properties of the generated surfaces are computed and shown to agree with the specified properties. The surface height density function of the generated surface is found to depend upon the number of filter weights chosen and the density function of input random numbers. The results are pertinent to the study of electromagnetic-wave scattering by rough surfaces