顾及森林散射体非随机朝向/非均匀分布的长波PolInSAR林下地形反演

So far, the scattering model based polarimetric SAR interferometry (PolInSAR) underlying topography mapping and forest height inversion assume that the forest is a random homogeneous volume and the SAR signal with specific polarization channel does not contain surface scattering contribution. However, in the vision of long-wavelength SAR, the forest scatterers are dominated by large branches, and the spatial orientation of the scatterers is non-random, as well as the spatial position distribution is heterogeneous. Moreover, due to strong penetrability, the SAR signals of all polarization channels contain significant ground surface echoes. In addition, by simplifying the polarimetric information, the traditional scattering model based inversion uses PolInSAR complex coherence as observation and the inversion process relies on the simple combination of multi-baseline PolInSAR data, making it difficult to fully and accurately utilize multi-baseline PolInSAR polarimetric information and polarimetric interferometry information as well as hard to fully consider the scattering characteristics of long-wavelength SAR in forest areas. Therefore, this project intends to investigate the polarimetric SAR scattering model considering the strong penetration and non-random orientation of the scatterers. Based on this, the PolInSAR scattering model considering the vertically heterogeneous distribution of forest scatterers is established. Finally, we plan to develop the complex adjustment methods based on multi-baseline PolInSAR coherency matrix for estimating the underlying topography, in order to form underlying topography mapping method suitable for long-wavelength PolInSAR data.

目前，基于散射模型的极化干涉SAR（PolInSAR）林下地形及森林高度反演研究多假定森林为随机均匀体，且存在特定极化方式的SAR信号不包含地表回波。但是，长波SAR视野下，森林散射体以大型枝干为主，散射体空间朝向非随机、空间位置分布非均匀，且由于强穿透性，所有极化方式的SAR信号均包含显著地表回波。此外，传统散射模型反演通过对极化散射信息简化以PolInSAR复相干性作为观测量，解算过程依赖多基线PolInSAR数据简单组合，难以充分准确利用多基线PolInSAR极化散射信息及极化干涉信息全面考虑长波SAR森林区散射特性。鉴于此，本项目拟研究考虑强穿透及散射体非随机朝向的极化SAR散射模型；在此基础之上，建立顾及森林散射体非均匀分布的极化干涉SAR散射模型；最后，发展多基线PolInSAR相干矩阵反演林下地形的复数平差方法，形成适用于长波PolInSAR数据的林下地形测绘方法。

目前，基于散射模型的极化干涉SAR（PolInSAR）林下地形及森林高度反演研究多假定森林为随机均匀体，且存在特定极化方式的SAR信号不包含地表回波。但是，长波SAR视野下，森林散射体以大型枝干为主，散射体空间朝向非随机、空间位置分布非均匀，且由于强穿透性，所有极化方式的SAR信号均包含显著地表回波。此外，传统散射模型反演通过对极化散射信息简化以PolInSAR复相干性作为观测量，解算过程依赖多基线PolInSAR数据简单组合，难以充分准确利用多基线PolInSAR极化散射信息及极化干涉信息全面考虑长波SAR森林区散射特性。鉴于此，本项目研究考虑强穿透及散射体非随机朝向的极化SAR散射模型；在此基础之上，建立了顾及森林散射体非均匀分布的极化干涉SAR散射模型；最后，发展多基线PolInSAR协同反演林下地形的复数平差方法，形成适用于长波PolInSAR数据的林下地形测绘方法。项目实施过程中发表论文27篇、授权发明专利5项。项目研究成果丰富了极化SAR散射理论、将平差理论由实数域扩展至复数域，相关理论成果：为国产长波PolInSAR系统实施林下地形测绘提供技术参考；支撑了“透视测绘”的定义，为穿透测绘理论与方法建立奠定了基础；作为理论基础，支撑无人机载L波段极化干涉SAR林下地形测绘系统的研制。