论文部分内容阅读
现有简缩极化(Compact Polarimetry)SAR图像H/α经验特征空间存在两个问题:一是没有考虑简缩极化模式下的散射熵普遍高于全极化模式;二是在散射机制重叠区域,简缩极化H/α空间的分类能力较弱,尤其是多次散射。针对以上问题首先定量分析了DCP模式简缩极化SAR的散射角与全极化SAR数据散射角之间的关系,且在对7组不同传感器的SAR数据分析的基础上,提出了散射熵的替代参数ED,基于Monte Carlo模拟实验得到了H/α分解的各参数(熵H、平均散射角α和替代参数ED)分解的稳态条件;然后通过统计各散射机制在ED/α分布的密度空间,提出了一种新的简缩极化SAR图像ED/α特征空间。实验结果表明:替代参数ED与全极化熵具有良好相关性,而且ED/α特征空间提高了散射机制分类的精度。
There are two problems in the existing empirical feature space of compact Polarimetry SAR images: one is that the scattering entropy in the simplified polarization mode is not considered generally higher than the total polarization mode; the other is in the overlapping area of the scattering mechanism, Reduced polarization H / α space classification ability is weak, especially the multiple scattering. In order to solve the above problems, the relationship between the scattering angle of DCP mode reduced polarization SAR and the scattering angle of full-polarized SAR data is quantitatively analyzed. Based on the SAR data analysis of seven different sensors, the scattering entropy is proposed Parameter ED, the steady-state conditions for the decomposition of each parameter of H / α decomposition (entropy H, average scattering angle α and substitution parameter ED) are obtained based on the Monte Carlo simulation experiment. Then by statistical scattering mechanisms in the ED / α density space , A new ED / α feature space of reduced-reduced polarimetric SAR images is proposed. The experimental results show that there is a good correlation between the ED and the total polarization entropy, and the ED / α feature space improves the classification accuracy of the scattering mechanism.