Parameter Estimation of Heavy-Tailed Rayleigh Prior Model from Observed SAR Amplitude Images

We propose a new method for estimating the parameters of heavy-tailed Rayleigh prior model from the observed synthetic aperture radar (SAR) amplitude images. First, we change the multiplicative SAR image model into an additive one using logarithmic transformation. Next, we derive the expectation and variance of the log-transformed image using negative-order moments concept. Finally, we obtain the above quantities for two kinds of SAR amplitude images: the square-root of intensity image and the multi-look averaged amplitude image. Particularly for the latter, we derive the closed-form expressions for such expectation and variance based on the Gaussian approximation of the log-transformed speckle. Monte Carlo simulation results demonstrate that the proposed estimators, which are easy to implement with the analytical expressions, are efficient for the parameter estimation of the heavy-tailed Rayleigh prior model from the observed image.