Scattering from non-Gaussian randomly rough surfaces-cylindrical case

A numerical method is developed to simulate electromagnetic wave scattering from computer-generated two-dimensional randomly rough surfaces. The rough surface generated for scattering simulation is specified only up to the second moment statistics, i.e. the height distribution and the autocorrelation function. The coherent and noncoherent scattering from four different types of random surfaces is examined. The four different types of surfaces are: Gaussian distributed heights and Gaussian correlation, Gaussian distributed height and non-Gaussian correlation, modified exponential distributed height and non-Gaussian correlation, modified exponential distributed height and Gaussian correlation, modified exponential distributed height and non-Gaussian correlation surfaces. It is shown by simulation that the dominating factor in coherent scattering is the surface height density and the autocorrelation can cause a higher order effect