基于动态平差理论的InSAR三维时序地表形变估计方法研究

Limited by the temporal discrepancies among different sensors and different tracks SAR data, InSAR basically can only measure one/two-dimensional displacement time series or three-dimensional instantaneous displacements/displacement velocities. This limitation greatly degrades the ability of InSAR in the monitoring geo-hazards. In order to accurately monitor geo-hazards in four dimension (i.e., three-dimension in space and one-dimension in time), in this project we introduce the dynamic adjustment theory to develop a novel method of estimating three-dimensional surface displacement time series by integrating multi-temporal, multi-sensor and multi-track InSAR measurements. In particular, we first develop the method of estimating one/two-dimensional displacement time series with InSAR based on the displacement parameter optimization. The observation model of estimating three-dimensional displacement time series with InSAR is then constructed according to the dynamic characteristics of geo-hazards. Third, we study the state model and estimation principle of estimating three-dimensional displacement time series with InSAR on the basis of the dynamic adjustment theory. Finally, the developed method is applied to monitor and analysis the three-dimensional surface displacement time series and their patterns in the Sichuan-Yunan region, China. The outputs of the project are expected to improve the temporal resolution and accuracy of InSAR three-dimensional displacement estimations, and strength the international influence of China in this research field. In addition, the results of the Sichuan-Yunan region will provide scientific evidences for the early-warning and prevention of geo-hazards and for the sustainable utilization of underground resources.

受限于不同平台、轨道SAR数据的时态差异，InSAR技术目前基本只能用于一维/二维时序形变或瞬态三维形变/三维形变速率监测，使其在面对复杂的地质灾害时受到较大局限。本项目创新性地引入动态平差理论，研究一种融合多时相、多平台和多轨道InSAR测量值估计三维时序地表形变的理论与方法，实现地质灾害的四维（空间三维+时间维）高精度监测。具体的，本项目首先发展基于形变参数优化的InSAR一维/二维时序形变估计方法，然后建立顾及地质灾害动力学特征的InSAR三维时序形变估计的观测模型，进而研究基于动态平差理论的InSAR三维时序形变估计的状态模型和估计准则，最后开展川滇典型区域的三维时序地表形变监测及规律分析。项目成果将提高InSAR三维形变估计的时间分辨率和精度，提升我国在该领域的国际影响力；而且可以获取川滇典型区域地质灾害的时空规律，为区域内地质灾害预警和防治、地下资源可持续利用提供科学依据。

星载雷达干涉测量（InSAR）技术是近年来迅速发展起来的一种空间大地测量手段，可实现全天候、全天时、大范围、高精度的地表形变测量。融合不同平台、不同轨道的SAR数据，可以将InSAR的传统雷达视线方向上的一维形变测量值拓展至三维。然而，受限于多源SAR数据的时态差异，现有研究方法基本集中在一维/二维时序形变或瞬时三维形变/三维形变速率监测，难以获取可靠的三维时序形变结果，进而无法准确地分析和反演地表形变背后的地球物理机制和人类工程活动。本项目则针对上述问题展开了四方面的技术攻关，具体包括（1）基于时序形变模型库与假设检验理论，提出了一种InSAR一维时序形变自适应估计方法，有效降低了传统方法因形变模型不准确引起的一维时序形变参数估计误差；（2）构建了基于地下弹性半空间理论的地下流体变化与地表三维形变之间的观测模型，使得仅利用单个成像几何的InSAR观测数据即可实现地下流体变化引起的三维地表形变；（3）发展了基于卡尔曼滤波的InSAR三维形变序列动态估计模型，通过速度位移模型有效克服了不同平台间InSAR数据时态差异带来的参数求解秩亏问题，同时在InSAR大数据背景下，通过动态平差技术，可显著提高InSAR数据处理效率；（4）开展了大量的模拟数据实验，以及青海涩北气田、新西兰地震、九寨沟地震、三峡新铺滑坡等真实数据实验。结果表明，本项目提出的模型方法可有效获取地表三维时序形变结果及相应的地球物理参数，从而为区域内地质灾害的预警和防治、地下流体和矿产资源的可持续利用提供重要的科学依据。本项目共发表SCI论文18篇、CSCD论文1篇；申请国家发明专利8项，其中6项已授权。