Parameter Estimation Of Sea Clutter Amplitude Distribution And Target Cell Collection Method

In recent years, moving target detection in sea clutter has been a hot spot in radar community. Sea cultter is referred to as the back scattering of the sea surface under the exposure of radar, whose non-stationary and non-Gaussian characteristics lead to the main obstacle to detect targets on sea surface. Therefore, establishing the suitable clutter model in the current circumstances of sea surface is the foundation and key to mitigating sea clutter and extracting target returns.Firstly, the characteristics of statistics and Doppler spectral of sea clutter are researched. Several common amplitude distribution models of sea clutter are mainly introduced, including Rayleigh distribution, the Log-Normal distribution, the Weibull distribution, the K distribution and Pareto distribution. Simultaneously, methods of generating their random series are given. Moreover, the mean and fluctuation properties of Doppler spectral of sea clutter are analyzed and discussed by real sea clutter datasets.Secondly, the moment method and the bi-percentile method are utilized to estimate the parameter estimate for the Pareto amplitude model of sea clutter. Experimental results using real sea clutter datasets show that the amplitude model based on bi-percentile method is fitter to the characteristics of sea clutter in comparison with the method of moment due to the natural ability of the bi-percentile method to resist abnormal values.Thirdly, considering the fact that the noise component is contained in the real sea clutter data, the model with noise component and the model without noise component in K distribution are respectively modeled and estimated. The results by real data show that the model considering noise component in K distribution is closer to the amplitude characteristic of real sea clutter, which indicates that it is necessary and practical to take account the noise component into the model of real sea data.Finally, a new approach of target aggregation is proposed due to the problem of distributed targets in distance and direction under the pattern of low resolution. For the large size of sea surface or the leak of side beam from strong target returns, the energy of signal returns will spread into the ambient direction and range cells, resulting in their probability of detection.Therefore, a method of target aggregation based on relative power and phase linear degree is proposed, which enables to precisely extract position information of targets and auxiliary information of the size of targets. Besides, its effectiveness and practicability are confirmed by real sea clutter datasets.