基于InSAR技术的三峡库区巫山-奉节段潜在滑坡识别

以三峡库区滑坡灾害多发的巫山-奉节段为研究区,基于InSAR技术提出了一种潜在滑坡早期识别方法。该方法首先选取2017年3月-2018年11月共48景Sentinel-1 A SAR升轨数据,利用永久散射体干涉测量(PS-InSAR)技术提取出相对稳定的散射体作为地面控制点(GCP),并引入到小基线集干涉测量(SBAS-InSAR)技术流程中获取长时间序列的地表形变信息;然后统计分析该区域历史滑坡的孕灾因子规律,结合区域SAR数据的地形可视性分析筛选出孕育滑坡概率大于50%的形变点,并对这些形变点采用Anselin Local Moran’s I指数进行空间域异常值分析和聚类处理,获得置信水平不低于95%的高异常值区;最后结合区域地形地貌特征和高分辨率光学影像识别出高概率的潜在滑坡范围。该方法具有可行性,可用于指导滑坡灾害预测预警和野外地质灾害的实地调查。

Identification of the Potential Landslide in Wushan-Fengjie in the Three Gorges Reservoir Area Based on InSAR Technology

This paper selectes Wushan-Fengjie as a research area,where landslide disasters occur frequently in Three Gorges reservoir area.The paper proposes a method to early identify potential landslide based on InSAR technology and applies the data of 48 uplifting sentinel-1 A SAR images from March 2017 to November 2018 to extracting relatively stable scatterers by permanent scatterer technology(PS-InSAR).The paper also introduces these relatively stable scatterers as ground control points(GCPs) into the small baseline(SBAS-InSAR) technical flow to obtain long-term sequence surface deformation information.The paper analyzes the pattern of causing factors of historical landslides in this region, studies the visibility of regional SAR data to screen out the probability of causing landslides which is greater than 50%.Outlier analysis and clustering methods(Anselin Local Moran’s I)are adopted to analyze these points and a high outlier area with a confidence level of no less than 95% is obtained.With the regional geomorphic features and high resolution optical images,the paper identifies potential landslide areas with high probability.The adopted method is feasible and can be used to guide field investigation.