Sea Clutter Characteristics

In this paper, characteristics of sea clutter are analyzed from respects of statistical properties, correlation and non-stationary feature. It is important for sea targets detection.Firstly, statistical properties of sea clutter are anaylzed. In this section, four kinds of statistical distribution are introduced, which are Rayleigh, Lon-normal, Weibull, and K distribution. And we analyze the influence of the parameters of the distributions,and then estimate the value of the parameters by some estimating methods.At last, the statistical distribution fitting is analyzed with measured data.Secondly, correlation is analyzed in this section. Time correlation is analyzed with the measured data. From composite model perspective, sea clutter can be divided into two parts, speckle and texture. The relevant time of speckle and texture is analyzed and simulated.Thirdly, non-stationary and target detection are discussed in this paper. It is showed that the Doppler Frequency and texture are changing with time and that they are related, and it can be concluded that the sea is non-stationary. And then the measured data difference is fitted by Tsallis distribution so that sea clutter data is non-extensive. Nelder-Mead is adopted to estimate the parameters'value of the Tsallis distribution.and the result shows that the parameters is sensitive to the sea targets. On this basis, considering the parameters'sensitivity, joint parameter object detection method is discussed, which is very useful to detect the sea targets. At last, BP neural network is applied to detect the sea targets, and it is effective.