基于时序InSAR的采空区地表形变监测研究

It is the basis for stability evaluation and safety utilization that accurate monitoring the ground deformation above the abandoned mining area. Multi-temporal InSAR (MT-InSAR) has been proven a powerful tool that can monitor ground deformation. However, there are two limitations in mapping mining-induced deformation when using mutli-temporal InSAR. One limitation is that the density of coherent targets, such as PS and conventional DS, is low due to the dense vegetation and scarce buildings. Another limitation is that the deformation model adopted by multi-temporal is inconsistent with the law of surface movement caused by mining. In order to overcome these limitations, the new type DS with multiple scatters (termed as M-DS) and the robust solution of deformation estimation based on the theory of mining deformation will be researched. Firstly, a method used for identifying the M-DS and extracting its phase based on the statistical analysis of multiple scattering mechanisms will be proposed to increase the density of coherent targets in abandoned mining area. Secondly, the solution of different coherent targets combined with PS, DS and M-DS, the method of deformation estimation based on L1 norm and the atmospheric phase compensation based on statistics of Weather Research and Forecasting (WRF) are studied to solve the problems introduced by error propagation, outliers and atmospheric disturbance. Finally, a stable and robust MT-InSAR that is suitable for the geographical environment and deformation characteristics of abandoned mining area will be established, and which can provide theoretical and technical support for the stability evaluation and safety utilization of abandoned mining area.

采空区地表形变的准确监测是地面建设场地稳定性评价及土地安全利用的基础。时序InSAR是目前国内外广泛应用的地表形变监测方法，但用于建筑物少、相干性差、非线性形变显著的采空区时，存在PS和DS目标稀少、形变解算模型不适用等缺陷。针对上述问题，本项目分析研究多散射特性DS（M-DS）的散射统计规律，构建M-DS识别模型、算法和准则，提高采空区的相干目标密度；研究联合PS、DS和M-DS多相干目标的解算策略、基于L1范数的形变估计方法、基于数字天气预测模型统计规律的大气相位补偿方法，解决误差传播控制、离群点影响和大气扰动问题，建立稳定、鲁棒的形变解算模型。在此基础上，形成适用于采空区地理环境和形变特点的时序InSAR处理方法，为采空区地表形变监测、稳定性评价以及土地安全利用提供理论和技术支撑。

采空区地表形变的准确监测是地面建设场地稳定性评价及土地安全利用的基础。合成孔径雷达干涉（InSAR）技术具有时空分辨率高、覆盖范围广、历史存档数据量大等优点，逐渐成为大地测量中的有力工具之一。但用于建筑物少、植被多、非线性形变显著的采空区时，InSAR技术存在相干性差、相干点稀少、形变解算模型不适用等缺陷。. 针对上述问题，本研究基于分布式散射体（DS）的协方差矩阵构建了多散射特性DS（M-DS）的识别模型和相位估计方法，提高了矿区的相干点密度；通过补偿大气相位、采用同质滤波技术提高相干质量，提高了时序InSAR在矿区的形变解算精度以及模型的稳定性；结合概率积分法模型设计了InSAR矿区三维位移监测预测方法。. 利用项目提出的方法和模型在巨野、薛湖和章丘矿区进行了数据处理、形变机制分析、三维位移预测以及地表稳定性评估，为采空区地表形变监测、稳定性评价以及土地安全利用提供理论和技术支撑。尤其是在巨野矿区发现一个值得进一步关注和研究的现象：距离开采区约4公里远处的地表也出现了沉降，而根据沉降的量值与空间分布形态推知该沉降很可能是远处地下煤炭开采时的抽排深层地下水所致。这一发现可以对划定重点建/构筑物的禁采线提供有益指导。