多角度极化SAR图像散射特征建模及其应用

Multi-Aspect PolSAR Scattering Signatures Modeling and Application

  • 摘要: 本文将基于条带观测模式的极化SAR散射模型拓展至方位向多角度观测模式,基于典型极化散射类型组合提出一种非各向同性散射特征模型。该模型参数纬度多且随方位向观测变化,需要替代性方法提取多角度极化散射特征。首先,采用基于Wishart分布的统计量对非各向同性散射中心进行检测,并逐像素生成基于散射特征差异的新序列图像。其次,以新序列图像作为处理对象,提取极化似然比序列、子孔径角度序列、极化熵—似然比序列、极化散射角—似然比序列、极化各向异性度—似然比序列。最后,集成特征序列编码及支持向量基(SVM)方法进行分类。通过机载P波段极化SAR开展360°观测试验,验证了方法的有效性并揭示出在地物分类方面的应用潜力。

     

    Abstract: A new polarimetric multi-aspect synthetic aperture radar scattering model was developed in this paper. The new model is highlyl and non-analytic. Therefore, an equivalent computable algorithm was developed. First, we quantified and ranked media polarimetric scattering dissimilarity over all aspects. Then, we proposed the polarimetric likelihood ratio, the sub-aperture aspect, the polarimetric entropy, the polarimetric scattering angle and the polarimetric anisotropic signatures. At last, a SVM classification scheme was used for validation and analyze.

     

/

返回文章
返回