复杂山区滑坡形变监测的相干散射体InSAR方法

Spaceborne InSAR technology has the advantages of wide coverage and low cost in landslide deformation monitoring. However, in complicated mountainous areas with vegetation cover and steep terrain, conventional time series InSAR methods still have many problems, such as sparse measurement points and tropospheric atmospheric delay disturbances, etc., which hinder the effective extraction and high-precision measurement of landslide deformation. Hence, this project will study the coherent scatterer InSAR method for landslide monitoring in complex mountainous areas, which combines the persistent scatterers and distributed scatterers. Firstly, the homogeneous pixels are accurately detected using both amplitude and phase information. Secondly, Then, the precision of optimal phase extraction is improved by optimizing the complex coherent matrix estimation, and the processing efficiency of massive radar data is improved by introducing the sequential processing method. Thirdly, we fuse the delays predicted from various numerical meteorological models to effectively correct the tropospheric delay in steep terrains. Finally, we carry out cross validation and accuracy assessment of the proposed method in typical landslide experimental areas. It is expected to enhance the deformation monitoring capability of spaceborne InSAR technology in complicated mountainous areas, and provide a reliable technical means for the monitoring, prevention and control of landslide geological disasters in China.

星载InSAR技术在滑坡形变监测中具有覆盖范围广、监测成本低等优势，然而，在植被覆盖和地形陡峭等复杂山区环境下，常规时间序列InSAR方法仍然存在诸多问题，例如测量点稀少、对流层大气延迟扰动等，阻碍了滑坡形变信息的有效提取和高精度测量。针对以上问题，本项目将深入开展针对复杂山区滑坡形变监测的相干散射体InSAR方法研究，对永久散射体和分布式散射体进行联合处理和分析。首先，提出利用幅度和相位信息准确探测同质像元；然后，通过优化估计复数相干矩阵来改善最优相位提取的精度，并引入序贯处理的思想来提高海量雷达数据的处理效率；其次，发展多种数值气象模型大气延迟的加权融合算法，有效改正地形陡峭山区的对流层大气延迟；最后，在典型滑坡实验区对相干散射体InSAR方法进行交叉验证和精度评估。通过本项目的研究，有望提升星载InSAR技术在复杂山区的形变测量能力，为我国滑坡地质灾害监测与防治提供可靠的技术手段。

星载InSAR技术在滑坡形变监测中具有覆盖范围广、监测成本低等优势，然而，在植被覆盖和地形陡峭等复杂山区环境下，常规时间序列InSAR方法仍然存在诸多问题，例如测量点稀少、对流层大气延迟扰动等，阻碍了滑坡形变信息的有效提取和高精度测量。针对以上问题，本项目深入开展了针对复杂山区滑坡形变监测的相干散射体InSAR方法研究。第一，提出了联合SAR影像幅度和相位信息构建复数协方差矩阵，基于广义似然比框架检验同质像元，并研究了基于张量分解的多时相极化SAR同质像元识别方法；第二，为了提升复杂山区最优相位提取精度，研究了各种基于复数相干矩阵优化估计的最优相位提取方法，并提出了一种序贯估计算法提升数据处理效率；第三，针对复杂山区大气延迟扰动问题，提出了垂直分层延迟迭代线性估计模型和融合气象模型方法，针对大气延迟的空间异质性变化问题，提出了自适应融合多源气象模型和多时相移动窗口线性迭代模型；第四，在四川省丹巴县，将本项目的相干散射体InSAR方法与其它典型时序InSAR方法进行对比，并经过外部GNSS观测数据验证了所提方法的有效性和可靠性；最后，在四川、贵州等地质灾害多发省份，开展了InSAR滑坡地质灾害识别与监测应用工作，为防灾减灾工作提供重要数据支持。此外，借助本项目支持，研发了一套运行于Web端的海量InSAR点云时序形变可视化和解译平台。通过本项目的研究工作，提升了星载InSAR技术在复杂山区的形变测量能力，可为我国滑坡地质灾害监测与防治提供可靠的技术手段。