Parameter Estimation Of Sea Clutter Amplitude Distribution Based On Compound Gaussian Model

In recent years,the state of the ocean to explore more and more attention,and ocean ocean currents itself has a very complex nature,in the radar to detect the target on the sea,the radar electromagnetic waves will inevitably produce a variety of complex backscatter with the sea,The formation of high strength,complex sea clutter,the target detection,positioning and tracking had a great hindrance.Especially in the high-resolution radar to work with low wiping angle,there will be a sea peak phenomenon,and the target characteristics are very similar,prone to false alarm.Modeling the sea clutter using the appropriate mathematical model is an effective means to eliminate the influence of sea clutter.The accurate estimation of the parameters of the distribution model is a necessary prerequisite for the analysis and suppression of sea clutter,effectively extracting the target from the complex clutter background.Firstly,this paper models the sea clutter from the statistical point of view,and introduces the amplitude distribution model of the sea clutter under different conditions: Rayleigh distribution,logarithmic normal distribution,Weibull distribution,generalized Pareto distribution,Inverse Gaussian distribution and K distribution model.And gives the method of random sequence generation.The information of the sea clutter in the frequency domain is also an important feature of the sea clutter.This paper analyzes the Doppler spectral properties of the sea clutter from the frequency domain and the IPIX radar data,which is complementary to the time domain analysis.Secondly,for the generalized Pareto distribution model,the moment estimation method and the double quantile estimation method of generalized Pareto distribution are reviewed.A new method for estimating the accuracy and estimation accuracy of the anti-anomaly scattering unit is proposed for the shortcomings of the moment estimation method and the double-point estimation method.,The two-point estimation and the truncation moment estimation method are used to estimate the parameters of the simulated clutter data and the measured clutter data.The results show that the cut-off moment estimation has better performance than the moment estimation and the double-point estimation,and can be used to estimate the parameters of sea clutter in many cases.Finally,based on the K-distribution model,the composite K-distribution model with additional Gaussian white noise is used to model the sea clutter under the assumption that the additive thermal noise component exists in the measured sea clutter data.Fractional Moment Estimation Method.And the K-distribution model and the K-distribution model of additional noise are used to model the simulation data and the measured data respectively.The results show that,especially in the case of low noise ratio,considering the noise situation more in line with the actual clutter environment.The results show that the fractional moment estimation method has better estimation accuracy.