植被覆盖地表电磁模型与土壤水分反演研究

The research objective of this project is to establish a physical scattering / radiation model for rough surface and for vegetation canopy with high accuracy and a wide region of validity. Then this physical model is used for developing parametric models, and for entropy-alpha decomposition. These outcomes will in turn be used for soil moisture retrieval from bare soil and small-to-moderately vegetated land, and for the best combination of frequencies/polarizations/incidence angles for an accurate inversion. It is hoped that through this project, the understanding can be deepened of the complicated interaction of electromagnetic waves with complex terrain, of the impact of physical models on the decomposition characteristics, of the parametric formalism and performance in a parameterization procedure. Our ultimate scientific goal of the project and the follow-up study is to provide some fresh ideas and approaches to the difficult problem of soil moisture inversion of vegetated land.

本项目的研究目标是建立粗糙地表和植被的精度高、适用范围广的物理散射/辐射模型，并进行相应的参数化模型建模，以及进行Entropy-Alpha分解，研究裸露地表及中小生物量植被覆盖地表条件下的土壤水分反演算法，研究数据源的最佳频段/极化/角度组合方式，以期获得较高的土壤水分反演精度。通过本项目的研究，加深对电磁波与复杂地表交互、物理模型对极化分解特性的影响以及参数化模型中参数化模式及性能的影响等诸多重要基础应用问题的认识。我们的最终科学目标是通过本项目及后续研究，能为解决土壤水分反演精度受植被影响大、随场景波动大这一国际难点和热点问题提供新的思路和方法。

在本课题支持下，建立了粗糙地表和植被的精度高、适用范围广的物理散射/辐射模型，并进行相应的参数化模型建模和初步的机器学习表征，并提出了对Durden等人极化分解理论的完善，研究了裸露地表及中小生物量植被覆盖地表条件下的土壤水分反演算法，获得了项目预期的土壤水分反演精度。在知名国际期刊（IEEE T-GRS，IEEE GRSL，IEEE T-AP，IEEE JSTARS)上已发表6篇标注论文，另有三篇在修。通过本项目及后续研究，有望为解决土壤水分反演精度受植被影响大、随场景波动大这一国际难点和热点问题提供新的思路和方法。.