基于高分辨率时间序列InSAR 数据分析的桥梁形变监测

Bridges are important infrastructure elements and the continuous monitoring of the stability and the safety of bridges is of utmost importance. Typically, the stability and the deformation of bridges is monitored with traditional surveying methods and some experiments are carried out with continuous GPS monitoring. Deformation measurement with Permanent Scatterer Interferometry(PSI) is an effective tool for precisely measuring the urban subsidence in an extend area. Using PSI for the monitoring of the bridge stability would provide a precise and.cost-effective way.Unfortunately, because of the strong vibrations of large bridges,this is diffcult for standard PSI techniques. Using PSI for the monitoring of large bridges requires an extension of the technique,which will be applied when necessary. The main idea is to keep the standard PSI method whenever possible,because it is a proven technique. To implemnet this successfully, the bridges need to be classified according to their bridge type and their coherence.Coherent bridges can be processed with standard PSI whith an additional focus on ghost target reduction and multi-scatterer processing. Partly-coherent and incoherent bridges are to be futher analyzed and classified according to their bridges and need to be processed differently. The PS points or PS candidates on the bridges need to be grouped and classified. For large suspension bridges, we need to distinguish between points on the bridges and points on the cabloes or pillars. Based on the grouping of these points and the semanticof the points, an extension of the PSI method is to be developed. For bridges the PSI method need to be extended including an additional error source: bridge vibrations. This error is not corelated in time,because the temporal resolution is too low, but correlated in space. After successful semantic grouping of the points, this spatial relationship can be modeled and it is therefore possible to estimate the error caused by bridge vibrations for each PS point.

随着社会经济的快速发展，我国兴建的各类桥梁设施井喷式增加，给桥梁的健康监测和安全评估带来了严峻挑战，桥梁变形监测已成为公共安全、土木工程和对地观测领域日益关注的研究热点。雷达干涉测量技术（InSAR）可以大范围快速地提取地物目标的形变信息，在桥梁形变监测方面有着巨大的应用潜力。本项目拟针对时间序列InSAR技术在桥梁稳定性监测中的难点，重点研究从高分辨率时间序列InSAR 数据中提取桥梁形变信息的理论和方法。在研究不同类型桥梁在SAR 和InSAR 数据中散射特性的基础上，将散射特性和时间相干性关联起来分析，拓展时间序列相位信号模型，突破常规方法的局限性，面向不同类型桥梁提出自适应的形变监测方法，应用先进的卫星遥感技术实现对桥梁形变的便捷精细监测。为桥梁安全使用和养护管理提供可靠的决策依据，并为其他基础设施的安全预警提供新的技术途径。

随着社会经济的快速发展，各类桥梁设施井喷式增加，桥梁变形、垮塌等灾害的频繁发生，给桥梁的健康监测和安全评估带来了严峻挑战，桥梁变形监测已成为公共安全、土木工程和对地观测领域日益关注的研究热点。合成孔径雷达干涉测量技术（InSAR）可以大范围快速地提取地物目标的形变信息，具有人力物力成本低、全天候全天时、不影响桥梁正常运营等优势，在桥梁形变监测方面有着巨大的应用潜力。然而，由于现代桥梁结构复杂多样，结构层面的精细化测量给时间序列InSAR技术的应用带来新的挑战。. 本项目针对目前时间序列InSAR技术在桥梁安全监测中存在的点目标提取密度和精度低、形变模型单一、形变解译难度大等难点，重点研究从高分辨率时间序列InSAR 数据中准确提取桥梁形变信息的理论和方法。在研究SAR影像中桥梁散射特性的基础上，综合分析桥梁的散射信号类型、结构相干特征、结构语义信息等进行点目标的匹配和筛选，提高结构点目标的识别精度；考虑桥梁结构特征与温度形变的分布规律，突破常规线性形变模型的局限，拓展时间序列相位信号模型；将形变量与温度变化进行建模关联分析，提取并分离桥梁的温度形变，提高桥梁形变测量的精度；基于结构特征和形变分布对桥梁进行形变关键点识别和全桥形变特征提取；基于对桥梁点目标的分类分析，探究桥梁不同结构组件上经验性的形变机理，探究不同类型桥梁的形变特征/机理与其特定结构/材料特性之间的潜在关联，并将提出的时间序列InSAR方法应用于高分辨率TerraSAR-X、Cosmo-SkyMed以及中分辨率Sentinel-1等SAR影像，实现了对不同类型桥梁形变的便捷精细监测。. 项目的实施挖掘了时间序列InSAR技术在桥梁结构安全评估方面的应用潜力，可以为桥梁的安全使用和日常养护管理提供科学可靠的决策依据，服务于桥梁形变所带来的一系列重大公共安全保障工作。