Clutter Suppression Method Based On Gaussian Mixed Model

Radar uses the reflection of electromagnetic waves from objects such as targets to detect the position of the target.However,the radar working environment is complex and there are many complex objects around the target.These objects also reflect electromagnetic waves,which are received by radar as a clutter signal.Therefore,the clutter suppression technique is one of the important technologies in the signal processing.Clutter suppression is the key technology to detect small and weak targets in clutter background.Clutter suppression technology means that radar can keep the target as much as possible and improve the detection performance of the target under the premise of suppressing the clutter around the target.At present,Moving Target Indication,Moving Target Detection,clutter map and constant false alarm are the most commonly used techniques to suppression clutter and detect targets,which are widely used in modern radar systems.However,when the velocity of target is close to the velocity of clutter or the Doppler distribution of clutter is wider,the traditional clutter suppression technique can't suppress clutter well,thus affecting target detection.In one resolution unit,the clutter is composed of one or more scatterers with random phase and amplitude.Therefore,in general,the probability density function can be used to describe the statistical characteristics of the clutter.With the continuous development of modern science and technology,a large number of clutter empirical models have been proposed,such as Gauss distribution model,Weibull distribution model and lognormal distribution model.In this thesis,the problem of clutter suppression is studied from the point of view of statistical distribution.Based on the idea of classification,a method of clutter suppression based on Gaussian mixed model is proposed.The theoretical basis of this method is that any probability density function can be fitted by multiple Gaussian distribution models,and the target and clutter amplitudes in radar echo data are not satisfied with the same distribution model.In this method,the Gaussian mixed model is first used to fit the statistical clutter distribution,and then the Gaussian mixed model is used to classify the echo data received by the radar,so as to suppress the clutter and detect the target.Finally the new method is applied to the clutter suppression processing of radar received measured data,and the target signal in the echo is detected,which verifies the feasibility and realizability of the clutter suppression method proposed in this thesis.The method proposed in this thesis can suppress most of the clutter in the Doppler and distance dimensional extended distribution.It has the advantages of good generality and high accuracy.It solves the problem that the current clutter suppression technology has a bad effect on the wide spectrum of the clutter suppression and the rejection of the clutter around the target.