基于光学影像的滑坡位移提取方法及应用

Method and application of landslide displacement extraction based on optical image

为实现通过遥感手段对滑坡灾情进行评价、对滑坡滑动性质做出判断的目的,本文以不同时间的Sentinel-2和UAV影像为数据源,采用光学遥感影像配准与关联(COSI-Corr)方法提取江顶崖滑坡的水平位移。对比分析了以同源和多源影像为数据源的计算结果,探究COSI-Corr对多源影像数据的兼容性。同源数据的试验结果表明,2018年江顶崖滑坡最大位移为140 m,平均位移为20.6 m。滑坡上部位移大于下部位移,即反映出该滑坡为推移式滑动。此外,结合降雨资料分析了滑坡的诱因。以多源影像为数据源,提取出了正确的滑动特征,表明了COSI-Corr支持多源影像融合关联。根据研究成果可判断灾情规模,为灾后工程治理提供参考,同时COSI-Corr也拓宽了UAV影像的应用范围。

In order to realize the purpose of using remote sensing to evaluate the landslide disaster and judge the landslide nature,this paper uses Sentinel-2 and UAV images at different time as the data source,and uses co-registration of optically sensed images and correlation(COSI-Corr)method to extract the horizontal displacement of the Jiangdingya landslide.This paper compares and analyzes the calculation results of homologous and multi-source images as data sources,and explores the compatibility of COSI-Corr with multi-source image data.The test results of the same source data show that the maximum displacement of the Jiangdingya landslide in 2018 is 140 m,and the average displacement is 20.6 m.The displacement of the upper part of the landslide is greater than that of the lower part,which indicates that the landslide is a push type of sliding.In addition,combined with rainfall data,the causes of landslides are analyzed.Using multi-source images as the data source,the correct sliding features are extracted,indicating that COSI-Corr supports multi-source image fusion association.The research results can judge the scale of the disaster and provide a reference for post-disaster engineering management.At the same time,COSI-Corr also broadens the application range of UAV images.