Predictive density of millimeter-wave backscattering based on Gamma mixture model

In this paper, we investigate a predictive density distribution of millimeter-wave (MMW) backscattering from terrain for the purpose of the estimation of MMW clutter distribution. In order to derive the density, the product model represented by speckle and texture is applied, where the distributions of the speckle and texture are modelled as the Gaussian and the Gamma mixture models, respectively. For the estimation of the mixture parameters, the expectation-maximization technique is applied. The applicability of the predictive density is demonstrated with experiment data at 94GHz band.