极化SAR目标散射特性分析与应用

发布时间：2018-04-13 23:38

本文选题：合成孔径雷达 + 极化　；参考：《清华大学》2014年博士论文

【摘要】：极化合成孔径雷达是极化、合成孔径、雷达三大技术的结合，是遥感的重要手段。极化SAR图像包含丰富的目标散射信息，如何充分利用这些信息是图像解译的关键。本文从四个方面对极化SAR目标散射特性进行了分析，，主要创新点如下： 1）通过提取最近的相干目标来对目标散射特性进行描述。从极化合成理论出发，定义了Kennaugh矩阵的范数。一个Kennaugh矩阵分解成两部分之和。其中一部分对应于一个相干目标，另一部分是个剩余项，其Kennaugh矩阵的范数达到最小。从功率保持和极化特性保持两方面，讨论了所提取相干目标的性质。 2）通过去定向操作对目标散射特性进行描述。定向角是极化SAR目标的一个重要参数，出于不同的应用背景，定向角有不同的定义方法。对8种定向角的不同定义进行了归纳总结，指出了其中的内在联系。讨论了极化SAR图像目标去定向的性质。分析了去定向前和去定向后二次散射系数的变化趋势。分别以这两个指标对极化SAR数据进行分类比较，结果表明后者能较好地区分三种常见的散射目标。 3）利用相似性参数对目标散射特性进行描述。在舰船检测中，通过计算目标与三种标准散射体之间的相似性参数，发现目标与杂波散射特性的差异。基于这种差异，对极化交叉熵的定义进行了改进。结合功率值，提出了功率极化交叉熵的概念。新参数能反映目标与杂波的功率、极化性质的不同。实验验证了该参数用于CFAR舰船检测的有效性。 4）提出了一种多波段联合表征目标散射特性的方法。对不同波段相干矩阵进行了独立性验证，通过特征值分解的方法提取出主散射成份，对不同波段的散射相关进行估计，将多波段的相干矩阵映射成一个新矩阵。该矩阵可以描述目标在多波段下的散射特征。经过Boxcar滤波，并对滤波后的数据进行Wishart监督分类。实验验证了所提分类方法具有简单易操作和准确率高等特点。
[Abstract]:Polarimetric synthetic Aperture Radar (ISAR) is an important means of remote sensing.Polarimetric SAR images contain abundant target scattering information, and how to make full use of these information is the key to image interpretation.In this paper, the scattering characteristics of polarized SAR targets are analyzed from four aspects. The main innovations are as follows:1) the scattering characteristics of the target are described by extracting the nearest coherent target.Based on polarization synthesis theory, the norm of Kennaugh matrix is defined.A Kennaugh matrix is decomposed into the sum of two parts.One part corresponds to a coherent target, the other is a residual term, and the norm of the Kennaugh matrix reaches the minimum.The properties of the extracted coherent targets are discussed in terms of power retention and polarization characteristic retention.2) the scattering characteristics of the target are described by the de-directional operation.Orientation angle is an important parameter of polarized SAR target.The different definitions of eight directional angles are summarized and their internal relations are pointed out.The properties of target de-orientation in polarized SAR images are discussed.The variation trend of the secondary scattering coefficient before and after de-orientation is analyzed.The polarimetric SAR data are classified and compared with each other, and the results show that the latter can distinguish three common scattering targets well.3) the scattering characteristics of the target are described by similarity parameters.In ship detection, by calculating the similarity parameters between target and three kinds of standard scatterers, the difference of scattering characteristics between target and clutter is found.Based on this difference, the definition of polarization cross-entropy is improved.Combined with the power value, the concept of cross-entropy of power polarization is proposed.The new parameters can reflect the difference of power and polarization between target and clutter.The effectiveness of this parameter in CFAR ship detection is verified by experiments.4) A multi-band joint method is proposed to characterize the scattering characteristics of the target.The independence of the coherent matrix in different bands is verified. The principal scattering component is extracted by eigenvalue decomposition, and the scattering correlation of different bands is estimated, and the multi-band coherent matrix is mapped to a new matrix.The matrix can describe the scattering characteristics of the target in multi-band.After Boxcar filtering, the filtered data are classified by Wishart.The experimental results show that the proposed classification method is simple and easy to operate and has high accuracy.
【学位授予单位】：清华大学
【学位级别】：博士
【学位授予年份】：2014
【分类号】：TN958

【参考文献】