海杂波中目标检测的时频方法研究

发布时间：2018-01-17 03:32

本文关键词：海杂波中目标检测的时频方法研究　出处：《西安电子科技大学》2014年硕士论文　论文类型：学位论文

更多相关文章： 海杂波 功率谱密度 时频特性 FRFT

【摘要】：随着对地高分辨观测技术的发展,海杂波的背景中目标的识别和检测越来越成为热门研究话题。海面目标的回波,主要是来自海面背景本身的杂波和目标的散射回波。要实现雷达对海面目标的检测,就要区分这两种散射回波的特征。因此,论文基于加拿大McMaster大学IPIX雷达实测数据,利用时频分析方法,研究背景海杂波和目标的散射回波的特性。论文的主要工作如下：1、介绍了两种典型的雷达海杂波数据的产生方法,即无记忆非线性变换法(ZMNL)和球不变随机过程法(SIRP)。重点用ZMNL法产生了符合瑞利分布、对数正态分布和韦伯儿分布的海杂波。用SIRP方法产生了符合相关K分布的海杂波,并分析了两种杂波的产生方法的优缺点。2、分析了海杂波的功率谱特性。研究海杂波功率谱的估计方法有经典的周期图法和现代谱估计法,用这两种方法分析了实测海杂波数据的功率谱,结果表明不同极化和海情的杂波的功率谱存在差异,同时发现现代谱分析方法比经典的周期图法取得更加平滑和清晰的效果。然后用相关函数法分析了海杂波的时间相关性,研究表明海杂波的时间相关性都有一个共同的特点就是经历过一个去相关时间,然后是一个比较长的衰减期直到相关函数值几乎为O。然后对海杂波的空间相关性作了分析。3、研究了实测海杂波数据的统计特性。列举了瑞利分布、对数正态分布、韦伯儿分布和K分布这四种统计分布模型。根据实测海杂波数据,用理论分布模型和实际分布进行了对比研究,结果表明相关K分布可以很好地拟合实测海杂波数据。4、描述了背景海杂波和带目标的海杂波的时频分析方法。包括短时傅里叶变换法(STFT),维格纳分布法(Wigner-Ville)法,Wigner-Hough变换法,分数阶傅里叶变换法(FRFT)。然后用实测海杂波数据对以上方法的有效性进行了验证和对比,发现短时傅里叶变换能够得到清晰的时频分布特性。然后用Matlab仿真实测海杂波数据的维格纳分布、平滑维格纳分布和平滑伪维格纳分布,仿真结果表明平滑伪维格纳分布能够有效克服交叉干扰项对维格纳分布带来的影响,取得较好的时频分析效果。同时通过研究背景海杂波数据和带目标的杂波数据可以发现维格纳分布可以识别海面慢速运动的目标,并且在二维的时频图上得到直线带状结构,而Wigner-Hough变换可以识别这种直线带状结构,通过坐标变换可以得到非常清晰的直线尖峰,从而进一步验证了Wigner-Ville方法对海面目标识别的应用。另外,还确定了背景海杂波和带目标的杂波在经过分数阶傅里叶变换后,在最佳变换区域内得到不同的分形特征,为海面目标识别提供了依据。
[Abstract]:With the development of high-resolution observation technology, target recognition and detection in sea clutter background has become a hot topic. It is mainly clutter from sea background and scattering echo from target. In order to detect sea surface target by radar, it is necessary to distinguish the characteristics of these two kinds of scattering echo. Based on the measured data of IPIX radar at McMaster University in Canada, the characteristics of background sea clutter and scattering echo of target are studied by using time-frequency analysis method. The main work of this paper is as follows: 1. Two typical methods for generating radar sea clutter data are introduced. That is, the memoryless nonlinear transformation method (ZMNL) and the spherical invariant stochastic process method (SIRP). The emphasis is on the ZMNL method, which is consistent with the Rayleigh distribution. The sea clutter with logarithmic normal distribution and Weber distribution is generated by SIRP method, and the advantages and disadvantages of the two methods are analyzed. The power spectrum characteristics of sea clutter are analyzed. The classical periodic chart method and modern spectrum estimation method are used to estimate the power spectrum of sea clutter data, and the power spectrum of measured sea clutter data is analyzed by these two methods. The results show that there are differences in the power spectrum of the clutter with different polarization and sea conditions. At the same time, it is found that the modern spectral analysis method is more smooth and clear than the classical periodic chart method, and then the correlation function method is used to analyze the temporal correlation of sea clutter. The research shows that the temporal correlation of sea clutter has a common characteristic, that is, it has experienced a de-correlation time. Then a long attenuation period until the correlation function value is almost O. then the spatial correlation of sea clutter is analyzed. 3. The statistical characteristics of the measured sea clutter data are studied and the Rayleigh distribution is listed. The four statistical distribution models of logarithmic normal distribution, Weber distribution and K distribution are compared and studied by using the theoretical distribution model and the actual distribution according to the measured sea clutter data. The results show that the correlation K distribution can fit the measured sea clutter data. 4. The time-frequency analysis method of background sea clutter and sea clutter with target is described, including STFT (short time Fourier transform). Wigner-Hough transform is a Wigner-Ville-based distribution method. The fractional Fourier transform method is used to verify and compare the effectiveness of the above method with the measured sea clutter data. It is found that the short-time Fourier transform can obtain a clear time-frequency distribution, and then the Wigner distribution, smooth Wigner distribution and smooth pseudo-Wigner distribution of sea clutter data measured by Matlab are simulated. Simulation results show that smooth pseudo-Wigner distribution can effectively overcome the influence of cross-interference on Wigner distribution. At the same time, by studying background sea clutter data and clutter with target data, we can find that Wigner distribution can identify slow moving target. And the linear band structure can be obtained on the two-dimensional time-frequency map, and the Wigner-Hough transform can identify the linear band structure, and can get a very clear straight line spike by coordinate transformation. The application of Wigner-Ville method to sea surface target recognition is further verified. In addition, the background sea clutter and the clutter with target are determined after fractional Fourier transform. Different fractal features are obtained in the optimal transform region, which provides a basis for the recognition of sea surface targets.
【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】：P715

【参考文献】