基于TOPS模式SAR影像的城市地表沿轨形变监测方法研究

Due to the lower azimuth resolution and smaller Doppler bandwidth of Sentinel-1 TOPS data, the accuracy of along-track measurements is limited, which is the key issue in three-dimensional (3D) monitoring of city ground deformation. The city of Shenzhen in Guangdong Province is selected as the field site. This research aims to develop a method to measure along-track displacements from Sentinel-1 TOPS images with high accuracy. 1）Taking advantage of the large differences of Doppler centroid frequencies between consecutive bursts in a TOPS image, the burst overlap interferometry technique is used to measure along-track displacements within the burst overlap area. A fully non-local filtering method is applied to reduce the effects of spatial variation of displacement on the phase noise level. Time series inversion is used to further raise the accuracy. 2) In burst non-overlap region, to improve the accuracy of along-track measurements, we propose two modified Multiple Aperture Interferometry (MAI) methods: a) time series MAI approach based on Non-Local (NL) InSAR Filtering; b) time series MAI approach based on Dual-polarimetric SAR Interferometry, in order to reduce the interferometric phase noise, increase the coherence, and eventually improve the overall accuracy of along-track measurements. This project, if approved, will provide technical support on the wide-area monitoring of ground deformation in urban areas of South China, in the meantime enable the analysis of the major driving factors of ground displacement, and supply scientific evidence for city planning, construction and engineering management.

Sentinel-1 (S1) TOPS影像由于方位向分辨率降低、多普勒带宽减小，限制了沿轨形变测量精度，是制约着城市地表形变三维监测的关键问题。本研究选择广东省深圳市为典型样区，重点研究基于S1 TOPS影像的沿轨形变高精度测量方法：1）利用TOPS模式在相邻脉冲重叠区的多普勒中心频差较大的优势，采用脉冲重叠干涉技术测量沿轨形变，基于全局滤波减小空间异质性对相位噪声的影响，并结合时间序列反演获取脉冲重叠范围内的高精度沿轨形变信息。2）针对非脉冲重叠区域的沿轨形变测量精度低的问题，分别开展基于NL InSAR滤波优化、基于双极化干涉的时间序列多孔径沿轨干涉（MAI）方法研究，降低干涉相位噪声，提高相干性，旨在整体提高S1 TOPS数据的沿轨形变测量精度。本项目的开展将为华南地区大范围地表形变遥感监测提供方法支持，同时有助于分析形变的驱动因素，为城市规划、建设和工程管理提供科学支持。

城市地表形变对建筑物、地铁及其它大型基础设施有很大潜在危害，危及人民生命财产安全和城市的可持续发展。星载合成孔径雷达干涉测量（Interferometric Synthetic Aperture Radar，InSAR）具有全天时、全天候、覆盖范围广、穿透能力强的优势，目前普遍用于地表形变监测中。但是，目前InSAR观测技术对于卫星轨道方向的形变测量精度低，是制约着城市地表缓慢形变三维监测的关键问题。本研究选择广东省深圳市为典型样区，开展基于Sentinel-1 (S1) TOPS模式SAR影像和高分辨率COSMO-SkyMed (CSK) SAR影像的沿轨形变高精度反演方法研究：基于SAR同质点集进行相位滤波，在保持城市地表空间异质性的同时，抑制相位噪声。在SAR同质点统计判别的基础上进行干涉相干性估计和偏差纠正，实现了干涉相干性优化；在此基础上对原始StaMPS (Stanford Method for Persistent Scatterers)方法进行适应性改造，提出了SHP StaMPS方法，提高了卫星视线方向（LOS）形变的InSAR反演精度；针对研究区常年多云多雨、水汽时空变化大、强对流天气多的特点，提出了结合GACOS (Generic Atmospheric Correction Online Service, GACOS)大气产品和InSAR时空滤波的大气延迟相位精细校正方法，进一步提高了沿海城市LOS形变的InSAR测量精度；针对Sentinel-1由于方位向分辨率降低、多普勒带宽减小，导致沿轨形变测量精度低的问题，提出基于S1和CSK升降轨数据的跨平台三维形变反演方法，融合传统LOS时间序列InSAR测量、脉冲重叠干涉和多孔径干涉测量，结合物理模型正则化约束条件，显著提高了卫星轨道向形变的反演精度。