基于分数阶Fourier变换的雷达目标检测算法的研究

发布时间：2018-11-26 15:40

【摘要】：雷达目标检测算法在军用和民用等各领域都具有很高的应用意义,提高检测目标可靠性是雷达的重要作用之一。针对在海杂波背景中弱动目标的检测,传统的方法是从统计意义的角度对海杂波进行建模,实现雷达对目标有无的自动检测,然而在更加复杂的海杂波背景下,统计模型并不能完全体现海杂波的全部特征。在时频分析方法中,提取目标的频率信息可将目标回波近似为LFM信号,FRFT非常适合处理此类非平稳信号。因此,本文主要研究基于分数阶Fourier变换的雷达目标检测算法,结合分形理论,从而提高系统对动目标的检测能力。本文首先给出了 FRFT的定义和时频特性,以及用到的性质和特点,并介绍了一种快速离散计算方法。随后重点讨论了三种估计最优阶次的方法,特别是对分数阶域中心矩的方法进行详细的讨论,此方法可更快速有效地找到最佳旋转角,为检测做好前期工作。其次,利用FRFT针对海杂波和目标回波信号的能量集聚性不同来实现目标检测。其中先根据多普勒效应对雷达回波信号建模,介绍了实验用到的IPIX雷达数据和统计模型拟合分析,并针对不同统计模型下的海杂波进行检测验证。针对低信杂比情况下检测能力下降问题给出了移位对消改进方法,即将信号延迟前后在最佳旋转角时的模值相对消,从而改变信杂比,提高检测性能,其中重点讨论了延迟时间对检测的影响。最后,结合FRFT域分形差异对目标进行检测。首先需要对回波进行FRFT,在其最佳分数阶域上信号的信杂比得以改善,利用小波Hurst指数法计算分数阶域IPIX雷达数据有无目标单元时的Hurst指数,根据不同距离单元Hurst指数不同实现海杂波与目标回波的有效区分,其中着重分析了最佳选转角的选取问题。随后与时域分形差异检测方法进行比较,证明在低信杂比时经过FRFT预处理的检测性能比时域分形检测更高。
[Abstract]:Radar target detection algorithm has high application significance in military and civilian fields. Improving the reliability of target detection is one of the important functions of radar. For the detection of weak moving targets in sea clutter background, the traditional method is to model the sea clutter from the point of statistical significance, to realize the automatic detection of radar target, but in the more complex sea clutter background, The statistical model can not fully reflect all the characteristics of sea clutter. In the time-frequency analysis method, extracting the frequency information of the target can approximate the echo of the target to LFM signal, and FRFT is very suitable for dealing with this kind of non-stationary signal. Therefore, this paper mainly studies the radar target detection algorithm based on fractional Fourier transform, and combines fractal theory to improve the detection ability of the system. In this paper, the definition and time-frequency characteristics of FRFT are given, as well as the properties and characteristics used, and a fast discrete calculation method is introduced. Then three methods of estimating the optimal order are discussed in detail, especially the method of center moment in fractional order domain. This method can find the best rotation angle more quickly and effectively, and do the preliminary work well for detection. Secondly, FRFT is used to realize target detection by using different energy concentration of sea clutter and target echo signal. Firstly, the radar echo signal is modeled according to the Doppler effect, and the IPIX radar data and statistical model used in the experiment are introduced, and the sea clutter under different statistical models is detected and verified. An improved method of shift cancellation is presented to solve the problem of low signal-to-clutter ratio (LSRR), that is, the relative elimination of the modulus value before and after the signal delay at the optimal rotation angle, so that the signal-to-clutter ratio can be changed and the detection performance can be improved. The effect of delay time on detection is discussed. Finally, combining the fractal difference in FRFT domain, the target is detected. Firstly, it is necessary to improve the signal-to-clutter ratio of FRFT, signal in its optimal fractional order domain. The wavelet Hurst exponent method is used to calculate the Hurst exponent when there is a target unit in the fractional order IPIX radar data. According to the different Hurst exponents of different distance units, the effective distinction between sea clutter and target echo is realized, and the selection of optimal rotation angle is emphatically analyzed. Then compared with the time domain fractal difference detection method, it is proved that the detection performance of FRFT pretreatment is better than that of time domain fractal detection at low signal-to-clutter ratio.
【学位授予单位】：大连海事大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TN957.51

【相似文献】