基于重轨干涉SAR相位信息的土壤湿度变化建模及反演方法研究

Soil moisture is of great importance in human activities in social and economic aspects, which is one of the most proper research subjects in quantitative study tendency of microwave remote sensing techniques. The scattering modeling and moisture inversion approaches could be extended to other multi-layer medium, such as grassland, agriculture, forest and snow/ice etc. quantitative applications.This study follows the most recent research progress in soil moisture estimations using SAR techniques. Aimed at the difficulties in soil moisture change detection through repeat pass InSAR approach, it focuses to the phase information of data.The major contents include implementing the investigation in coherent scattering modeling of two-layer soil, developing the relevant inversion methods and flowcharts, and studying on the means of retaining or retrieving phase information in the inversion process. With the research on theory, methods and key techniques, the experiments would be carried out in controllable environment and also with spaceborne or airborne SAR data. This study would extent the application ranges of available InSAR data, meantime it could probably improve the quantitative inversion accuracy of soil moisture, and provide the techniques for optimizing the deformation measurements in those areas covered with soil surface.

土壤湿度对人类的社会和经济活动影响重大，是微波遥感向着定量化方向发展的最适合的研究对象之一，研究其散射模型、湿度反演的方法等，可以进一步拓展到其他的多层媒质，例如草地、农田和森林、冰雪等定量化应用中。本项目紧跟SAR技术土壤湿度的最新研究进展，针对重轨干涉SAR提取土壤湿度变化量研究中的难点问题，围绕数据的相位信息，开展双层媒质的土壤相干散射建模方法的研究，建立适用于土壤地表的双层媒质干涉相干散射模型，发展模型对应的反演方法和流程，并研究湿度变化反演过程中的相位信息保持及提取方法，完成理论方法研究及核心技术攻关，开展可控环境数据及实际星载、机载SAR数据的实验验证。本研究将拓展现有干涉数据的应用范围，同时为改善土壤湿度的定量化反演精度、优化土壤覆盖区域的形变监测结果提供技术基础。

土壤湿度已被全球气候观测系统列为关键气候变量，对人类的社会和经济活动影响重大。本项目依据土壤表层物理基础，建立了适用于土壤地表的双层媒质干涉相干散射模型，并在此基础上研究了相位三项组对土壤湿度反演的影响，并分析了滤波方法对相位信息的影响，利用星载SAR数据初步验证了干涉相位及相干性监测土壤湿度变化的可行性；接着以物理模型为基础并结合最新的数据驱动方法深度学习技术利用机载SAR数据来反演裸土地区及植被覆盖地区的土壤湿度，提高土壤湿度反演精度；最后对土壤介电模型、地形、大气变化等去相干因素的影响进行了分析。