基于时序差分干涉测量的黄土沟壑矿区地表沉降监测研究

发布时间：2018-05-24 22:00

本文选题：柠条塔矿区 + 地面沉降　；参考：《长安大学》2014年博士论文

【摘要】：合成孔径雷达干涉测量(SyntheticApertureRadar Interferometry，InSAR)及差分干涉测量(Differential InSAR,DInSAR)技术是近年来迅速发展起来的一种的空间对地观测技术，可全天候、大范围、快速地获取高精度的地表形变信息，在地球科学的诸多领域如地震、火山、矿区塌陷、地面沉降、滑坡、冰川移动等地质灾害监测中得到了广泛的应用。矿区地面沉降是一个复杂的地表形变过程，常规的地面沉降监测方法如水准与GPS等难以满足快速、大面积监测的需要。InSAR技术是有效的监测手段之一。为解决InSAR技术中时空等去相关因素的影响，近年来逐渐发展起来的基于相干点目标的时序差分干涉测量技术成了研究热点。但是针对研究区域的差异和数据集的特点，时序差分干涉技术的相关算法和环节还有待进一步优化。本文以陕北黄土沟壑区域的神木县柠条塔矿区为实验区，基于9幅ENVISAT/ASAR影像，利用时序差分干涉技术，，在干涉纹图集构建、相干点目标识别、三维相位解缠、地表沉降量提取、DEM残差估计、大气延迟分离等方面进行了深入分析与研究，提取了N1201工作面在2009.04-2010.09期间的地表沉降发育情况，结合开采情况与地面观测结果进行对比验证。本文的主要研究内容可分为以下几个部分： 1)在系统分析DInSAR基本原理、技术特点、误差来源和数据处理流程的基础上，分析了参考DEM误差、卫星轨道误差及地表形变对差分干涉相位的贡献程度，并推导了相应的比对公式。 2）相干点目标选取。分别从振幅、相位、振幅与相位信息相结合的角度，详细比较了基于幅度稳定性、相位稳定性、时序相关性及振幅离差阈值等多种点目标提取方法。针对SAR数据集特点，提出了基于时序相关系数阈值与振幅离差阈值并集的点目标识别策略，给出了相应模型。 3)地面沉降速率反演模型探讨。借助网络流的基本思想，采用最小费用网络流相位解缠模型，从时间维再到空间维的相位解缠，以时间维的累积相干系数和空间维的时间相干系数作为权重费用函数。利用相邻PS点之间的自相关特性建模，采用最大模型相关系数法、线性回归法对地表线性形变速率模型了进行探讨。同时，基于相位解缠结果和残余相位的特性，采用时间域高通滤波和空间域低通滤波分离出非线性形变分量和大气噪声分量。 4）地面沉降信息验证及分析。基于9幅ENVISAT/ASAR影像，利用时序差分干涉测量技术对柠条塔矿区地面沉降进行了实证研究，分离出了DEM高程偏差和大气噪声相位，提取了地表沉降发育情况，并结合开采情况和地面实测数据进行了对比分析。
[Abstract]:Synthetic Aperture Radar Interferometry (InSAR) and differential Interferometry (DInSAR) are rapidly developed space-to-earth observation techniques in recent years. It has been widely used in many fields of earth science, such as earthquake, volcano, mining area collapse, land subsidence, landslide, glacier movement and other geological hazards monitoring. Land subsidence in mining area is a complex process of surface deformation. Conventional monitoring methods such as leveling and GPS are difficult to meet the requirement of rapid monitoring. InSAR technology is one of the effective monitoring methods. In order to solve the influence of spatio-temporal and spatio-temporal de-correlation factors in InSAR technology, the time-series differential interferometry based on coherent point targets has become a hot research topic in recent years. But according to the difference of the region and the characteristics of the data set, the related algorithms and links of the temporal differential interferometry need to be further optimized. In this paper, the Caragana Tower Mining area of Shenmu County in the loess gully region of Northern Shaanxi is taken as the experimental area. Based on 9 ENVISAT/ASAR images, the interference pattern atlas is constructed by using time series differential interferometry, the coherent point target is identified, and the 3D phase unwrapping is carried out. The Dem residual error estimation and atmospheric delay separation of surface subsidence extraction are analyzed and studied in depth. The ground subsidence development of N1201 face during 2009.04-2010.09 is extracted, and the comparison between mining situation and surface observation results is carried out. The main contents of this paper can be divided into the following parts: 1) based on the systematic analysis of the basic principle, technical characteristics, error source and data processing flow of DInSAR, the contribution of reference DEM error, satellite orbit error and ground deformation to the differential interferometric phase is analyzed. The corresponding comparison formula is deduced. 2) selection of coherent point target. From the angle of amplitude, phase, amplitude and phase information, the extraction methods based on amplitude stability, phase stability, temporal correlation and amplitude deviation threshold are compared in detail. According to the characteristics of SAR data set, a point target recognition strategy based on the combination of temporal correlation coefficient threshold and amplitude deviation threshold is proposed, and the corresponding model is given. 3) the inversion model of land subsidence rate. Based on the basic idea of network flow, the phase unwrapping model of minimum cost network flow is adopted, from time dimension to spatial dimension, the cumulative coherence coefficient of time dimension and time coherence coefficient of space dimension are taken as weighted cost function. Based on the autocorrelation characteristics of the adjacent PS points, the linear deformation rate model of the surface is studied by using the maximum model correlation coefficient method and the linear regression method. At the same time, based on phase unwrapping results and residual phase characteristics, nonlinear deformation components and atmospheric noise components are separated by time-domain high-pass filtering and spatial low-pass filtering. 4) validation and analysis of land subsidence information. Based on 9 ENVISAT/ASAR images, the ground subsidence in Caragana Tower mining area was studied by time series differential interferometry. The DEM height deviation and atmospheric noise phase were separated, and the ground subsidence development was extracted. Combined with the mining situation and the measured data on the ground, a comparative analysis is carried out.
【学位授予单位】：长安大学
【学位级别】：博士
【学位授予年份】：2014
【分类号】：P642.26;P225.1

【参考文献】