Research On Improved Clutter Map Detection In Compound Gaussian Clutter

In recent years,with the application of detection in sea clutter background getting more and more attention in the radar field,moving target detection in sea clutter has been a hot spot.Because of the increase of radar resolution,the Rayleigh distribution model is no longer suitable for describing sea clutter.As a statistical model of semi-physical mechanism,the compound gaussian distribution can well describe the statistical properties of sea clutter.When the clutter characteristics varies seriously in the spatial domain,the detection perfor-mance of traditional methods will be greatly affeected,while the clutter map technology is an effective algorithm to detect the target.Therefore,this thesis focuses on the adaptive clutter map detection of moving targets in compound gaussian clutter background.A new plane clutter map algorithm based on cell median processing is also proposed.Finally,a new de-tection method based on clutter map technique and matched filtering technique is proposed in combination with the near optimal detector in the K distribution.The main research results of this thesis can be summarized as follows:At first,from the point of view of statistical modeling,the existing distributed model de-scribing the statistical characteristics of sea clutter amplitude of different resolution radars is reviewed.The amplitude distributions of sea clutter under the compound Gaussian model with Gamma texture and inverse Gamma texture are discussed in detail.Moreover,the cor-relation of sea clutter in the pulse domain,spatial domain and scanning domain is analyzed and discussed by real sea clutter datasets.Secondly,several classical non-coherent detection algorithms in spatial domain are reviewed.Two non-coherent detection algorithms which match the clutter characteristics are derived under the spatially-varying K distribution and the spatially-varying Pareto distribution clut-ter background.The CFAR properties of the algorithms are analyzed to be related to shape parameter,the number of reference cells,the false alarm rate and the number of integrated pulses.The Nitzberg clutter map technique and the cell average clutter map plane technique which detecting the target in the scanning time are discussed in detail.A plane clutter map detection based on cell median processing is proposed because there may exist abnormal points in the clutter environment.The detection performance of the proposed detector is verified by experiments based on simulation data.The experiment results show that the proposed method is robust to outliers.Finally,from the point of view of coherent detection,the Moving Target Detection technol-ogy and the detection method based on adaptive matching filtering are introduced in detail.According to the characteristics that the correlation of sea clutter changes slowly in the s-canning time,a new detection method combining the matched filtering technique and the multi-frame combined clutter map technology is proposed.In this method,the detection re-sult is used as a basis for determining,whether the next scanning data is updated,so that the"self-shadowing" problem caused by slows or stationary targets in the traditional clutter map technology is solved.The experimental results show that the detection performance of the proposed detector is better than that of the existing clutter map detection algorithms.