A Mixed Spectrum for Predicting Scattering Properties of Typical Rough Surface

It is instructive to exploit the electromagnetic scattering of random rough surface for inversion of surface parameters and high-resolution earth observation. The rough surface is usually modeled as a kind of rough surface power spectrum with gaussian distribution in height, and the correlation length varies with different functions, such as the gaussian and exponen-tial correlation function. However, for rough surfaces in natural scenarios, the common power spectrum often cannot accurately simulate the backscattering characteristic of the rough surface under the different incident angles. Therefore, in order to solve this problem, a two-parameter correlation function based on a combination of gaussian and exponential correlation functions is presented to better predict the electromagnetic scattering properties of typical rough surfaces. First, the mixed power spectrum expressions based on gaussian and exponential power spectrum are presented and the corresponding correlation lengths are given. Secondly, Advanced Integral Equation Model (AIEM) is introduced to calculate the electromagnetic scattering properties of the rough surface. Then, the pre-processing of the measured data and the selection range of the rough surface parameters are presented. Finally, Genetic Algorithm (GA) is used to search the optimal parameters of the mixed and exponential spectrum for fitting the measured data. Numerical results show that for the cement and pitch surface, the mixed spectrum can better predict the rough surface backscattering characteristics at different incident angles than the ex-ponential spectrum. For the rough surface such as sand table, the RMSE between the measured data and the prediction results of the mixed spectrum at X- and Ku-bands is smaller than that of the measured data and the exponential spectrum.