海杂波背景下的过采样MTD方法研究

发布时间：2018-02-28 04:27

本文关键词： 海杂波目标检测恒虚警过采样　出处：《西安电子科技大学》2014年硕士论文　论文类型：学位论文

【摘要】：随着科技的发展,人们对海洋的探索力度不断加大,这就使得海杂波在军事和民用领域都受到了广泛的关注。海杂波是指在海面目标检测过程中不希望探测,并且想要通过信号处理方法进行抑制的来自海表面的雷达回波。海杂波是一个非平稳的随机过程,这一特性构成了海面目标检测的主要障碍,而在海杂波背景下的目标检测问题中,海杂波特性的研究又是必不可少的。到目前为止,统计建模法仍然是描述海杂波特性的主要方法,海杂波的幅度分布模型就是基于该方法进行研究的。根据雷达入射角度、分辨率以及频谱宽度的不同,海杂波的幅度特性可以用瑞利分布、对数-正态分布、韦布尔分布以及K分布模型来描述。本文简要介绍了这几种分布模型,并给出了产生这几种分布信号的方法。除了幅度分布特性外,海杂波的多普勒谱特性在多普勒域的动目标检测中也起着非常重要的作用。本文通过一组实测数据计算了海杂波的多普勒谱,并对频谱偏移、带宽以及能量随着多普勒通道的波动特性进行了讨论和分析。其次,本文回顾了时域中的几种非相参目标检测方法,包括均值类恒虚警检测器和有序统计类恒虚警检测器,均值类恒虚警检测器包括单元平均恒虚警检测器、取大恒虚警检测器和取小恒虚警检测器。通过实测数据对这些检测器的检测性能进行比较和分析,实验结果表明在杂波环境比较平稳且在参考单元内没有强干扰信号的时候,单元平均恒虚警检测器有较好的检测效果,当参考单元内有强干扰信号的时候,有序统计类检测器的检测性能优于其他几种检测器。因为这几种检测器都只用到了回波的幅度信息,所以只有在目标回波非常强烈的时候才能进行有效的检测。最后,本文研究了海杂波背景下的动目标检测方法,该方法是通过对时域信号进行离散傅里叶变换,然后对各通道数据进行比率检测来实现。在检测过程中当进行临界采样傅里叶变换的时候,检测性能会有部分损失,这是由于多普勒通道中的傅里叶变换结果与多普勒导向矢量失配引起的。因此,本文提出了过采样傅里叶变换方法来弥补性能损失,并且判决门限采用依赖于幅度分布形状参数的方法获得。运用实测数据进行检测的时候,通过对检测性能和计算的复杂度进行折中考虑,发现过四采样能够达到比较好的效果。
[Abstract]:With the development of science and technology, the exploration of the sea has been intensified, which makes the sea clutter receive extensive attention in both military and civil fields. Sea clutter refers to the fact that the sea surface target detection process does not want to be detected. And the radar echo from the sea surface is suppressed by signal processing method. The sea clutter is a non-stationary random process, which constitutes the main obstacle of sea surface target detection. In the problem of target detection under sea clutter background, it is necessary to study the characteristics of sea clutter. Up to now, the statistical modeling method is still the main method to describe the characteristics of sea clutter. The amplitude distribution model of sea clutter is based on this method. According to the difference of radar incident angle, resolution and spectrum width, the amplitude characteristics of sea clutter can be obtained by Rayleigh distribution, logarithmic normal distribution. Weibull distribution and K distribution model are described. This paper briefly introduces these distribution models and gives the methods to produce these distributed signals. The Doppler spectrum characteristics of sea clutter also play an important role in the detection of moving targets in Doppler domain. In this paper, the Doppler spectrum of sea clutter is calculated by a set of measured data, and the spectrum is offset. The characteristics of bandwidth and energy fluctuation with Doppler channel are discussed and analyzed. Secondly, several non-coherent target detection methods in time domain are reviewed, including mean class CFAR detector and ordered statistical CFAR detector. The mean class CFAR detector includes unit average CFAR detector, large CFAR detector and small CFAR detector. The experimental results show that when the clutter environment is stable and there is no strong interference signal in the reference unit, the unit average CFAR detector has a better detection effect, and when there is a strong interference signal in the reference unit, The detection performance of ordered statistical detectors is superior to that of other detectors. Because these detectors only use the amplitude information of echo, they can only be detected effectively when the target echo is very strong. Finally, In this paper, the method of moving target detection under sea clutter background is studied. The method is based on discrete Fourier transform (DFT) of time domain signal. In the detection process, when the critical sampling Fourier transform is carried out, the detection performance will be partially lost. This is due to the mismatch between the Fourier transform results in the Doppler channel and the Doppler steering vector. Therefore, an oversampling Fourier transform method is proposed to compensate for the performance loss. And the decision threshold is obtained by the method of depending on the shape parameter of amplitude distribution. When the measured data is used to detect, the detection performance and computational complexity are considered in a compromise, and it is found that four sampling can achieve better results.
【学位授予单位】：西安电子科技大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：TN957.51

【相似文献】