基于ADS-InSAR的兰新高铁深层黄土隧道山体形变监测与反演

The high-speed railway from Lanzhou to Xinjiang is the very important railway in the northwest area in China and cross through the Huangtu highland area and Qilian Mountain area with various types of topography. In these areas there are many geological disasters, especially the landslide and surface cracking due to the mountain deformation, which greatly reduce the safety of high-speed railway operation. This project aims to effectively monitor the deformation of high-speed railway deep tunnels in these areas. The methodology involving in this project consists of four steps: 1) to develop a method to conduct the automated identification and detection of adaptive distributive objects; 2) to develop a model based on the optimized combination of baselines by the integration of the time series of SAR images collected in ascending and descending orbits to achieve the deformation information; 3) to conduct a continue and longtime monitoring by combing with GB-InSAR for deformation regions containing high-speed rail in typical mountain areas and correct the deformation data from InSAR in order to analyze the deformation evolution high-speed railway deep tunnels in these areas; 4) to perform the prediction for the deformation trends for the mountains using the environment index, and conduct the risk estimation and warning in order to provide the technical supports for the safe operation and maintain of high-speed railway.

兰新高铁作为我国“一带一路”战略计划的重要组成部分，穿越地形地貌复杂的黄土高原与祁连山区。脆弱的地质环境导致大范围深层黄土地质灾害频发，特别是山体形变所造成的滑坡、地表开裂等严重威胁高铁线路的安全运营。为进行有效的高铁深层黄土隧道山体形变监测，本项目提出自适应分布式目标差分干涉 (ADS-InSAR)方法提取山体形变信息：开展山体自适应分布式目标自动识别与探测方法的研究；联合使用卫星升降轨SAR影像时间序列建立基线组合优化的InSAR形变解算模型；结合地基InSAR对典型山体形变区实施长时序连续观测并对星载SAR形变数据予以校正，获得高铁沿线深层黄土隧道山体时序形变演变规律；基于环境影响因子分析预测山体形变趋势并开展线路安全风险评估与预警，为高铁线路的安全运营和维护提供可靠的技术支撑。

兰新高铁是我国西北高寒风沙区域修建的首条高速铁路，该条线路穿越地质与气象条件复杂的黄土地区和极端气候，其综合修建技术难度之高，堪称中国最复杂、最具挑战性的高铁。由于沿线区域复杂的地形地貌特点，兰新高铁隧道分布密集，这些隧道多建设于深层黄土区，地质环境极为脆弱，气候干燥且降雨集中，地质灾害频发，尤其以山体位移形变灾害(包括滑坡、地表开裂等)最为严重。.近年来，国内外诸多学者开展了基于多时相卫星SAR影像探测形变时空演变的研究，提出了相应的时序差分雷达干涉理论与方法，可统称为时序差分雷达干涉方法(Multi-temporal InSAR，MTI)。由于其具有观测范围大、形变测量精度高、自动化程度高等技术优势，已在大范围地表形变测量方面取得了较好的应用效果。.然而，MTI方法应用于山体形变(包括滑坡)监测的效果目前却不甚理想，主要包括：山体形变隐患点识别较为困难、山体位移形变的空间细节难以捕捉以及形变信息获取单一。最近研究发现，诸如岩屑区、裸露土坡等一些分布型目标(Distributed Scatterer, DS)仍然具有利用价值。通过分析DS能克服在黄土山体上无PS存在的问题，大大提高有效形变监测点的探测数量，从而获取高密度形变场。.为此，本项目针对深层黄土山体DS目标的散射特性及统计特征，联合升降轨Sentinel-1A数据，使用ADS-InSAR干涉方法提取山体位移形变信息。对典型深层黄土隧道(张家庄隧道)山体进行长时序连续监测，并基于研究区域地形地貌等环境影响因子，综合分析和预测山体的时序运动特征，开展深层黄土山体形变灾害的风险评估与预测预警，为深层黄土区高铁线路的运营和维护提供可靠的技术支撑。