| 基于方差分量估计的多源InSAR数据自适应融合形变测量 |
| |
| 引用本文： | 敖萌,张路,廖明生,张丽.基于方差分量估计的多源InSAR数据自适应融合形变测量[J].地球物理学报,2020,63(8):2901-2911. |
| |
| 作者姓名： | 敖萌 张路 廖明生 张丽 |
| |
| 作者单位： | 武汉大学测绘遥感信息工程国家重点实验室, 武汉 430079 |
| |
| 基金项目： | 国家重点研发计划（2019YFC1509201，2017YFB0502700），国家自然科学基金（41774006，41904001）资助. |
| |
| 摘 要： | 近年来,合成孔径雷达干涉测量(Interferometric Synthetic Aperture Radar,InSAR)技术在地面沉降监测方面展现了巨大的应用潜力,但受其重访周期和一维形变测量能力的限制,仅利用单一轨道卫星观测数据很难揭示真实的地表形变特征及其演化规律.随着在轨运行的SAR卫星系统不断增加,使得融合相同时间段内覆盖同一区域的多源多轨道InSAR数据成为可能.然而目前普遍采用的多源InSAR数据融合方法均为针对大尺度形变监测设计,或者忽略南北向形变甚至水平形变,容易造成误判.为此,本文对经典小基线集(Small Baseline Subset, SBAS)时序InSAR分析方法进行改进,在其形变反演模型中加入东西向和南北向形变参数,采用方差分量估计方法解算多源观测数据验后方差,通过迭代精化确定权重矩阵,从而获得形变参数的最优估值.使用美国南加州地区的ALOS PALSAR和ENVISAT ASAR数据开展实验,利用南加州综合GPS网(SCIGN)位于研究区域内的9个站点观测数据进行验证,结果表明本文方法得到的融合形变测量结果在垂直向上能够准确反映地表形变波动,周期性与GPS观测比较一致;同时,融合得到的三维形变场显示南加州洛杉矶地区存在不可忽略的水平形变,东西向形变测量精度略高于南北向.因此,基于方差分量估计的多源InSAR融合方法在提高形变测量时间序列连续性的同时,能够更准确地反演研究区域三维形变特征.
|
| 关 键 词： | InSAR 地面沉降 方差分量估计 时间序列 三维形变 |
| 收稿时间： | 2019-08-19 |
Deformation monitoring with adaptive integration of multi-source InSAR data based on variance component estimation |
| |
| AO Meng,ZHANG Lu,LIAO MingSheng,ZHANG Li.Deformation monitoring with adaptive integration of multi-source InSAR data based on variance component estimation[J].Chinese Journal of Geophysics,2020,63(8):2901-2911. |
| |
| Authors: | AO Meng ZHANG Lu LIAO MingSheng ZHANG Li |
| |
| Institution: | State Key Laboratory of Information Engineering in Surveying, Mapping, and Remote Sensing, Wuhan University, Wuhan 430079, China |
| |
| Abstract: | In recent years, Interferometric Synthetic Aperture Radar (InSAR) technology has shown great potentiality in land subsidence monitoring. However, due to the limitation of repeat cycle and one-dimensional line-of-sight (LOS) deformation measuring capability, it is difficult to use only one data stack collected by a single satellite to characterize the real surface deformation and evolution regularity. The continuously increasing number of SAR satellite systems makes it possible to fuse multi-source InSAR data in different bands and orbits with overlap in spatial and temporal domain. The integration methods currently used are more suitable for large-scale deformation monitoring, or neglecting the north-south deformation and even the horizontal deformation, which usually leads to misjudgment. Therefore, in this paper the Small Baseline Subset (SBAS) deformation inversion model is modified by incorporating horizontal deformation parameters in east-west and north-south directions. The posterior variance of multi-source data is determined by the Variance Component Estimation (VCE) method to iteratively refine the estimated weight matrices, leading to the optimal estimation of deformation parameters. Experiments were conducted with ALOS PALSAR and ENVISAT ASAR data over Southern California, USA. Measurements at 9 GPS stations located in the study area were used to perform validation, and the results show that the fused deformation measurement accurately reflect the surface deformation fluctuation in vertical direction and the periodicity is largely identical as GPS observations. While, the fused deformation field illustrates horizontal deformation trend in this area cannot be ignored, and the displacement measurement accuracy along east-west direction is better than that along north-south direction. Therefore, the proposed multi-source InSAR fusion method based on VCE can not only improve continuity of time series deformation monitoring, but can also characterize the 3-D deformation field of the study area accurately. |
| |
| Keywords: | InSAR Land subsidence Variance Component Estimation (VCE) Time series 3-D deformation |
| 本文献已被 CNKI 等数据库收录！ |
| 点击此处可从《地球物理学报》浏览原始摘要信息 |
| 点击此处可从《地球物理学报》下载免费的PDF全文 |
|
