基于Wigner高阶谱的水下简单形状目标回波处理

发布时间：2018-09-14 16:19

【摘要】：对于主动声呐目标检测与识别这一水声领域研究热点,如何从接收信号中检测到目标回波,并从回波中提取目标特征信息是其基础,也是其关键技术。根据目标回波与背景噪声在时频域上的分布特点以及目标回波亮点间的时频差异,人们更倾向于使用时频方法对该非平稳信号进行处理。随着人们对于高阶谱的兴趣的提升,高阶谱己经成为一种分析非高斯过程,提取信号更多谱信息的重要工具,但是它并不适用于非平稳过程。由此本文提出使用时变高阶谱——Wigner高阶谱来进行目标回波检测与亮点提取。本文根据Wigner高阶谱的定义,推导了 CW信号的Wigner双谱和Wigner三谱的基本表达式,仿真研究了 Wigner高阶谱在高斯白噪声背景下对CW信号和LFM信号的检测能力,并对其进行简要分析。研究了多分量CW信号和LFM信号的Wigner双谱和Wigner三谱的交叉项问题,给出不同信号形式下交叉项的不同特点,指出模糊域核函数交叉项抑制方法存在的不足。针对该缺点,根据形态学处理的基本算法及其作用,结合多分量Wigner双谱和Wigner三谱交叉项特点,本文提出了基于形态学处理的多分量信号Wigner高阶谱交叉项处理方法,并且仿真验证了该方法的有效性。此外,根据噪声与信号在Wigner高阶谱域的分布特点,指出该方法同样可以用于噪声的抑制处理。基于目标回波亮点模型,本文给出了典型水下目标模型回波在不同入射角度下的几何亮点的个数及相互间的时延差,并对其进行高阶时频特性分析。采用基于形态学交叉项处理方法,使用Wigner高阶谱进行目标回波亮点处理,通过仿真分析,给出了该方法的适用范围。最后根据实验数据处理结果,验证了目标回波亮点模型的正确性,验证了 Wigner高阶谱与形态学相结合的几何亮点提取方法的有效性,验证了上述方法在目标回波检测中的有效性。
[Abstract]:For active sonar target detection and recognition, which is a research hotspot in underwater acoustic field, how to detect the target echo from the received signal and extract the target characteristic information from the echo is the foundation and the key technology. According to the distribution characteristics of target echo and background noise in time-frequency domain and the time-frequency difference between the bright spots of target echo, people prefer to use time-frequency method to process the non-stationary signal. With the increasing interest in higher-order spectra, high-order spectrum has become an important tool for analyzing non-Gao Si processes and extracting more spectral information from signals, but it is not suitable for non-stationary processes. In this paper, a time-varying high-order spectrum, Wigner high-order spectrum, is proposed for target echo detection and bright spot extraction. Based on the definition of Wigner high-order spectrum, the basic expressions of Wigner bispectrum and Wigner trispectrum of CW signal are derived. The ability of Wigner high-order spectrum to detect CW signal and LFM signal under the background of Gao Si white noise is simulated and analyzed briefly. In this paper, the crossover of Wigner bispectrum and Wigner trispectrum of multicomponent CW signal and LFM signal is studied. The different characteristics of crossover terms in different signal forms are given, and the shortcomings of the method of suppressing crossover terms in fuzzy domain kernel function are pointed out. In view of this shortcoming, according to the basic algorithm of morphological processing and its function, combined with the characteristics of multi-component Wigner bispectrum and Wigner trispectral crossover, this paper proposes a multi-component signal Wigner high-order spectral crossover processing method based on morphological processing. The effectiveness of the method is verified by simulation. In addition, according to the distribution of noise and signal in Wigner higher-order spectral domain, it is pointed out that this method can also be used for noise suppression. Based on the target echo bright spot model, this paper presents the number of geometric highlights and the time delay difference between them under different incident angles of typical underwater target model, and analyzes their higher-order time-frequency characteristics. Based on morphological cross term processing method, Wigner high order spectrum is used to deal with the bright spot of target echo. Through simulation analysis, the application range of this method is given. Finally, according to the experimental data processing results, the correctness of the target echo bright spot model is verified, and the validity of the geometric bright spot extraction method combined with Wigner high order spectrum and morphology is verified. The effectiveness of the above method in target echo detection is verified.
【学位授予单位】：哈尔滨工程大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：TB56

【参考文献】