Statistical analysis of sea clutter with α-stable distributions

In this article, and to contribute to the study of the problem of maritime surveillance, we propose to characterize sea clutter using a statistical approach. In general, the statistical study of sea clutter requires to have significant material installed in a maritime environment, this to acquire a database of real signals broadcast by a sea surface. Therefore, in this paper, we choose to generate the sea surface by using the Elfouhaily sea spectrum. On the other hand, the Physical Optics approximation is used to calculate the diffusion coefficient of a sea surface. In this sense, a database of scattering coefficients is well established (for the fixed electromagnetic characteristics of the sea water), by varying the observation geometry, the polarization of the wave (at the emitter and at the receiver), the wind speed and wind direction. To analyze the distribution of electromagnetic scattering coefficients, we propose to use for the first time, the alpha-stable distribution, we compare the results obtained with those given by the two commonly used in statistical distributions: the Weibull and K law. The Kolmogorov-Smirnov evidence is used to show that the alpha-stable model is best suited to our database. Therefore, we look more specifically the influence of the polarization, the wind speed and wind direction on the parameters of alpha-stable distributions. The parameter delta particular position allows to discriminate the probability density of the parameter delta for some configurations.