大拖尾雷达杂波模型及其背景下的扩展目标检测方法研究

发布时间：2018-07-05 10:56

本文选题：高分辨雷达 + 扩展目标　；参考：《国防科学技术大学》2014年博士论文

【摘要】：随着雷达分辨率的提高,雷达目标特性与杂波统计特性均发生了深刻的变化,这对雷达目标检测问题提出了更高的要求。本文采用理论分析与实验验证相结合的研究方法,针对高分辨条件下扩展目标检测问题,系统研究了大拖尾雷达杂波模型及其参数估计、二维相关大拖尾分布杂波仿真以及基于一维距离像和SAR图像的扩展目标检测问题。第一章根据目前雷达技术发展的趋势和实际应用的需求,阐述了高分辨条件下扩展目标检测的背景和意义,简明扼要地总结了雷达杂波建模与仿真以及扩展目标检测的研究现状和发展趋势,并对本文的主要工作进行了概括。第二章研究了雷达杂波统计建模问题。首先对目前常见的雷达杂波模型进行了综述与分类,然后根据实际应用的需求,总结了一个性能优越的杂波模型应具有的特点,并通过分析对比,指出了基于乘积模型发展而来的KK分布与G0分布具有广泛的应用前景。第三章研究了大拖尾杂波模型的参数估计问题。首先综述了经典的参数估计方法,并通过理论分析与实验验证指出了各自的优缺点,然后提出了一种基于粒子群优化的参数估计方法,该方法将杂波数据统计直方图与杂波模型概率密度函数在部分采样点上的差异作为代价函数,通过粒子群优化搜索参数的最优值。然后分析了影响该方法参数估计精度的主要因素,最后通过仿真与实测杂波数据验证了该方法的优越性。第四章研究了二维相关大拖尾杂波的模拟问题。首先提出了利用分段模拟的方法改善非线性变换的精度,以及具有任意相关函数的二维相关高斯杂波的生成方法,然后研究了基于MNLT的二维相关KK分布与G0分布雷达杂波的仿真方法;推导了KK分布与G0分布的SIRP特征概率密度函数表达式,研究了基于SIRP的二维相关KK分布与G0分布雷达杂波的仿真方法。仿真结果表明,两种方法产生的二维相关大拖尾雷达杂波不论是幅度特性还是相关特性均满足设定的要求。第五章研究了大拖尾杂波背景下基于一维距离像的扩展目标检测问题。首先对检测问题进行了数学描述,阐述了一维扩展目标检测需要解决的主要问题,然后基于KK分布与G0分布两种大拖尾杂波模型提出了最优检测器、GLRT检测器以及OS-GLRT检测器等,并对这些检测算法进行了仿真与比较,指出OS-GLRT是一种比较实用的检测器,并在不同的条件下对该检测器进行了性能分析。第六章研究了大拖尾杂波背景下基于SAR图像的扩展目标检测问题。首先对SAR图像目标检测算法进行了综述与分类,然后重点分析了CFAR检测算法的研究方向,提出了一种基于KK分布的全局CFAR检测算法,并基于实测高分辨SAR图像与基于G0分布的全局检测算法以及基于自动筛选的CFAR检测算法进行了对比,说明了本文所提算法的有效性。第七章对全文工作进行了系统地总结,并给出了进一步研究的方向和建议。
[Abstract]:With the enhancement of radar resolution, the radar target characteristics and the clutter statistical characteristics have undergone profound changes, which put forward higher requirements for the radar target detection problem. In this paper, a combination of theoretical analysis and experimental verification is adopted to study the large trailing radar in high resolution conditions. Wave model and its parameter estimation, two dimensional large trailing distribution clutter simulation and extended target detection based on one dimension range image and SAR image. The first chapter expounds the background and significance of the detection of extended target under high resolution conditions according to the current development trend of radar technology and the requirement of practical application. The research status and development trend of the modeling and Simulation of the clutter, and the development trend of the extended target detection are summarized. The second chapter studies the problem of the radar clutter statistical modeling. Firstly, the common radar clutter models are summarized and classified, and then a superior performance is summarized according to the practical application requirements. In the third chapter, the parameter estimation problem of the large trailing clutter model is studied in the third chapter. First, the classical parameter estimation method is summarized, and the theoretical analysis and experimental verification are presented. In this method, a parameter estimation method based on particle swarm optimization is proposed. This method uses the difference between the statistical histogram of the clutter data and the clutter model probability density function at the partial sampling point as the cost function, and searches the optimal value of the parameters by the particle swarm optimization. Then, the estimation accuracy of the parameter estimation is analyzed. The main factor is to verify the superiority of the method by simulation and measured clutter data. In the fourth chapter, the simulation of two dimensional large trailing clutter is studied. First, the accuracy of the nonlinear transformation is improved by the method of piecewise simulation, and the generation method of two-dimensional related Gauss clutter with arbitrary correlation function is studied. The simulation method of two dimensional correlation KK distribution and G0 distributed radar clutter based on MNLT is studied. The expression of the probability density function of the SIRP feature of the KK distribution and the G0 distribution is derived, and the simulation method of the two-dimensional correlation KK distribution based on SIRP and the clutter of the G0 distributed radar is studied. The simulation results show that the two two dimensional correlation large trailing radar is produced by the simulation results. The fifth chapter studies the problem of extended target detection based on one dimension distance image in the background of large trailing clutter. Firstly, the mathematical description of the detection problem is carried out, and the main problems to be solved in one dimension extended target detection are discussed, and then two kinds of KK distribution and G0 distribution are based on the problem. The large trailing clutter model proposed the optimal detector, the GLRT detector and the OS-GLRT detector, and simulated and compared these detection algorithms. It was pointed out that OS-GLRT was a more practical detector and analyzed the performance of the detector under different conditions. The sixth chapter studied the SAR image in the background of large trailing clutter. First, we summarize and classify the SAR image target detection algorithm, then focus on the research direction of the CFAR detection algorithm, and propose a global CFAR detection algorithm based on KK distribution, and based on the measured high resolution SAR image and the global detection algorithm based on G0 distribution and the CFAR based on the automatic screening of CFAR. The detection algorithm is compared, and the effectiveness of the proposed algorithm is illustrated. The seventh chapter systematically summarizes the full text work, and gives the direction and suggestions for further research.
【学位授予单位】：国防科学技术大学
【学位级别】：博士
【学位授予年份】：2014
【分类号】：TN957.52

【参考文献】