【作者】 张弓；

【导师】 朱兆达；

【作者基本信息】 南京航空航天大学 ， 通信与信息系统， 2002， 博士

【摘要】 脉冲多普勒(PD)雷达是一种先进的全相参体制的雷达。经过近50年的发展，PD雷达技术已经发展的非常成熟。近年来这种雷达体制更多地与其他雷达体制兼容工作，是信号处理技术发展和应用的重要领域之一。本文是结合课题“某型机载PD雷达恒虚警检测性能研究”和“某型舰载三坐标雷达解模糊实现研究”进行的。论文对PD雷达信号处理中的若干关键问题进行了研究，主要内容包括：杂波跟踪技术、基于杂波跟踪的恒虚警(CFAR)检测和雷达目标解模糊处理。 第一章绪论，回顾了PD雷达技术的历史和发展，分析雷达自动检测和CFAR检测技术、PD雷达解模糊技术的现状。介绍了本文研究工作的目的和主要内容。 第二章讨论了机载PD雷达杂波模型建立与杂波仿真。介绍了一种机载PD雷达杂波仿真算法；对其提出了改进，增加了对高度线杂波和陆、海交界杂波的有效描述。通过仿真数据与实测数据比较，证明了该算法对机载PD雷达杂波仿真的有效性。 第三章介绍了用于杂波跟踪通道对杂波数据预处理的非线性滤波器设计。我们研究了相关的非线性滤波器，将K邻域滤波算法用于杂波预处理；并提出了一种改进的塔型形态滤波算法，用于杂波预处理，改善了杂波灰度图峰值信噪比，满足了杂波跟踪通道对杂波数据预处理的要求。 第四章研究了两种基于图像处理的杂波跟踪技术。模糊C均值(FCM)算法：将FCM算法应用于杂波跟踪，利用杂波在时间和空间的相关性，提出一种确定初始隶属度矩阵的方法，使FCM算法的迭代次数减少，降低了运算的开销；局部熵算法：结合机载PD雷达杂波分布特征和局部熵算法的特点，将局部熵算法用于杂波跟踪，由一次局部熵滤波、二次局部熵滤波与K邻域滤波组合，构成混合型滤波器，提出一种最大局部熵相似性判决准则，与混合型滤波器结合对杂波进行分割。两种跟踪方法在跟踪结果的形式、运算效率有所不同，可分别适用于二维和一维CFAR检测处理。对经过预处理后的仿真杂波数据采用上述两种算法进行分割处理，得到了满意的杂波跟踪结果。 第五章讨论了基于杂波跟踪的CFAR检测器。我们将杂波跟踪结果作为CFAR处理的背景信息，使杂波边缘估计与CFAR检测分离，提出了基于FCM算法和局部熵算法杂波跟踪的两种CFAR检测方法：杂波属性(clutter feature)，CF-CFAR检测器，和杂波结构元素(structuring element)，SE-CFAR检测器。并对其在杂波边缘处的检测性能进行了仿真分析和研究。在精确跟踪杂波的条件下，该检测方法在杂波边缘区与传统的CA、GO-CFAR方法比较具有良好的抗杂波边缘特性，接近理想检测性能。两种CFAR检测方法在参考单元结构、运算效率有所不同，可分别适用于二维和一维检测。 第六章研究了PD雷达解模糊处理。首先对聚类算法进行了研究和改进，在此基 张弓：机载PD雷达信号处理若干关键技术研究 础上提出了一种滑窗相关器算法，并对其性能进行了分析和仿真；在某型雷达低空对 海通道的数据处理中，基于并行DSP系统实现了该滑窗相关器算法。另外还介绍了 我们将实数遗传算法用于解模糊处理的研究工作。结合解模糊原理，设计了相应的适 应度函数，通过仿真证明：该算法使得解模糊正确率明显提高。 第七章结束语，对全文工作进行了总结，并指出下一步需要继续研究的问题。更多还原

【Abstract】 PD radar is a kind of advanced pulse radar which using coherent processing. It provides an active field for development of the signal processing techniques. This dissertation is based on the subject" Research on constant false alarm rate (CFAR) detection for one certain type of airborne PD radar " and " Research on ambiguity resolution in range and velocity for one certain type of shipborne radar". In this dissertation several key problems about signal processing for PD radar have been studied. The main content includes: Clutter tracking technique, CFAR detector based on clutter tracking and ambiguity resolution in range and velocity etc.In chapter one, the origin and development of PD radar technique are reviewed. The automatic radar detection and CFAR processing methods, ambiguity resolution for PD radar are summarized. The research contents and objectives of the dissertation are given.In chapter two, airborne PD radar clutter models are discussed. A kind of airborne PD radar clutter emulation algorithm is introduced and improved. The expression enables the accurate and efficient simulation of radar sidelobe, main lobe, altitude-line clutter and coastline clutter. Comparing emulated and real data, we proved the validity of this emulation to airborne PD radar clutter.In chapter three, we discuss non-linear filter designing used for pre-processing clutter in clutter tracking channel. We apply K-nearest neighborhood (K-nn) algorithm to pre-process clutter and propose a modified scheme based on pyramidal morphology algorithm. The modified algorithm does better than the original algorithm on the Peak Signal-to-noise Ratio (PSNR) of clutter gray image and follows the requirement on pre-processing clutter in clutter tracking channel.In Chapter four, the clutter tracking technique based on image processing is proposed. We apply Fuzzy C-means Clustering (FCM) algorithm to the clutter tracking. According to the clutter dependence in the time and space, we propose a method of defining the initial fuzzy membership function. This method is verified to be successful in reducing the computation burden using simulated data. We apply Local Entropy algorithm to the clutter tracking. The hybrid filter, which is composed of K-nn filter and local entropy filter, has been used to detect clutter edge. We propose the judging criteria of similarity to Maximum Local Entropy that is used to segment the simulated clutter data after local entropy filter. The FCM algorithm and Local Entropy algorithm have difference in the form of output, efficiency and suitable CFAR detectors. Those algorithms have been verified to be successful in several typical simulations.In Chapter five, we introduce the research on CFAR detection based on the clutter tracking. Regarding the result of clutter tracking as the background information we propose clutter feature (CF) -CFAR detector and structuring element (SE)-CFAR detector. They are based on different results of clutter tracking. They have different reference cell structure and are suitable for one or two-dimensional detection respectively. Their detection performance in clutter edge region are analyzed and simulated by the MonteCarlo method. They have very good behaviors compared with CA, GO-CFAR detectors at clutter edges.In Chapter six, the research on ambiguity resolution in range and velocity in PD radar is discussed. First we study and improve cluster algorithm, put forward a sliding window correlator algorithm. This algorithm’s performance is analyzed and emulated. Then we introduce the design and realization of this algorithm in one certain type of ship-borne radar by parallel DSP. In the last part, we introduce real-valued genetic algorithm applied to ambiguity resolution. Based on the ambiguity resolution principle, we design the fitness function. A good result in the probability of ambiguity resolution using real-valued genetic algorithm is get.In chapter seven, the research results are concluded. The further developments and the questions needing continuing studying are p更多还原