摘要
森林区域林下地形及树高的反演是极化干涉雷达的一个重要应用。该文首先对极化干涉SAR数据的相干区域进行建模及运用最优正规矩阵近似干涉互相关矩阵,得到白化正规干涉互相关矩阵。白化正规干涉互相关矩阵的相干区域为一条直线,任意求得两个不同极化状态下的相干系数进行直线拟合,完成地表的估计,再结合体散射去相干与树高之间的关系,运用查表方法完成树高的估计。该方法回避了传统方法中求解所有极化状态下的相干系数估计及相干区域边缘提取的步骤,在简化参数反演提升估计效率的同时获得正确地表与树高估计,最后运用仿真数据完成算法有效性与可靠性的验证。
The inversion of topography and tree height in forest area is one of the most important applications in the Polarimetric SAR Interferometry （PolInSAR）. In this paper, the coherent region of the PolInSAR data is modeled and the best normal matrix is used to approximate the cross correlation matrix, further, the whitened interferometric cross-correlation matrix is obtained. The coherence region of the whitened interferometric cross- correlation matrix is a straight line. Two arbitrary coherences obtained under two different polarization states can be applied to fitting a straight line. Based on the fitting line, the topographical phase can be estimated successfully. Referring to the relationship between the volume scattering and the tree height, look-up table method is used to search the correct tree height. The proposed method can avoid the complex steps of the traditional method, which needs to solve all the coherences under different polarization states to obtain the edge of the coherent region. The proposed method simplifies the inversion procedure and improves the efficiency of inversion, meanwhile, achieves the correct topography as well as the tree height. Finally, the simulation data are applied to validating the validity and reliability of the proposed method.
出处
《电子与信息学报》
EI
CSCD
北大核心
2017年第5期1051-1057,共7页
Journal of Electronics & Information Technology
关键词
极化干涉合成孔径雷达
相干区域
最优正规矩阵近似
地形与树高反演
Polarimetric SAR Interferometry （PolInSAR）
Coherent region
Best normal matrix approximation
Topography and tree height inversion
