The Weak Target Detection Technology Under The Background Of Composite Sea Clutter

Target detection based on CFAR method in sea clutter has been a hot issue of radar detection technology. CFAR detection algorithm always uses a statistical distribution model of clutter to design the detection threshold adaptively, which remains relatively constant false-alarm probability. But when the real observation model is inconsistent with hypothesis statistical distribution model or there is an error in clutter parameter estimating, it is hard to guarantee the constant false alarm performance of CFAR detector, or even worse.Based on the research background of CFAR detection technology under complex sea clutter, the main contents of the thesis include: modeling and simulation of complex sea clutter, CFAR detection in sea clutter, CFAR technology research based on multi-frame detection etc. Thesis is summarized as follows:The second chapter firstly introduces common statistical distribution models in the sea clutter simulation, such as Log-Normal, Rayleigh, Weibull, etc. After analyzing the characteristics and scope of application, it carries on the simulation of the three models. In considering the time and space-based correlation, the chapter proposes a new kind of composite K distributed clutter generation based on ZMNL method. Simulation results show the effectiveness of the method, meanwhile provides some data support for the study of later algorithms. Finally, the measured data on the effect of several clutter model fitting comparison show that the composite K distribution model fits the measured sea clutter data most satisfactory.The third chapter mainly studies the constant false alarm detection algorithm of single frame observation distance-doppler figure. Firstly, the adaptive parameter CFAR detection algorithm model is introduced. Then it introduces the CFAR detection algorithm based on ordered data rate, and analyzes the performance. Finally, the high order fractal characteristics gap value is introduced ad the target judgment whether target is detected or not. Accordingly the paper puts forward a new detection algorithm based on the fractal characteristics of gap rate in RD image domain. The results of the simulation data and measured data for the three algorithms comparison test demonstrate the effectiveness of the algorithm.The fourth chapter studies CFAR technology based on multi-frames-detection under complex sea clutter background. Firstly, it proposes the multi-frame detection method based on continuous sequence of related frames and proves that the method can effectively reduce the false alarm rate. Then this paper introduces the process observation of Track-Before-Detection algorithm based on dynamic programming, and analyzes detection and tracking performance of DP-TBD and SOSCA-CFAR method. Finally an improved TBD algorithm based on Gaussian particle filter is put forward, which approximates the posterior distribution using a single Gaussian filter and uses sequential ratio test algorithm for multiple frame data accumulation. Measured data results show that the algorithm possesses good detection tracking performance.