Landslide Deformation Monitoring Based On Offset Tracking Of Optical Remote Sensing Images

As a global geological disaster,the landslide makes all parts of the country experience the harm caused by the landslide to varying degrees.In order to effectively prevent the occurrence of landslide disasters,scholars from all over the world use a variety of advanced remote sensing methods to carry out wide-area remote sensing detection,remote sensing deformation monitoring and remote sensing risk assessment of the landslide.In-depth research has further developed landslide identification and landslide detection technology,and the accuracy of surface deformation monitoring results has been effectively improved.Optical remote sensing images do not need to lay out points.They have the advantages of wide coverage of satellite signals,rich spectral information,and no interference,and can cope with large gradient surface deformation in various complex areas.With the increase of optical remote sensing satellites and the improvement of image quality,the use of optical remote sensing images to detect the surface deformation of geological disasters such as landslides shows great potential for development.How to use various types of images to monitor surface deformation and how to integrate multiple types of optical remote sensing images to improve the time resolution of surface deformation monitoring is becoming more and more important.This paper mainly analyzes and discusses the development of optical remote sensing system,the research status of landslide based on optical remote sensing and the development status,technical principle and data processing process of offset tracking technology.The deformation characteristics of Baige landslide in Jinsha River from 2015 to 2022 and the deformation characteristics of Slumgullion landslide in North America from 2016 to 2020 are obtained by optical image offset tracking technology.The main research results are as follows :1.Using 16 Sentinel-2 images,the displacement time series of Baige landslide in Jinsha River from 2015 to 2018 were inverted by COSI-Corr software and CARST software respectively.The maximum horizontal displacement of the slope before the occurrence of “10·10” Baige landslide was 31.69 m,and the feasibility of CARST software in landslide time series deformation analysis was confirmed by comparing and analyzing the two-dimensional deformation results and accuracy evaluation results.Combined with the influence of external factors such as rainfall,topography and geology,the spatial and temporal evolution process of Baige landslide in Jinsha River is analyzed.It is found that the spatial and temporal displacement of Baige landslide shows the deformation characteristics of “initial start→constant deformation→accelerated deformation ”.2.Using optical images and radar images to analyze the surface deformation of the “11·3” Baige secondary landslide from October 15,2018 to November 2,2018,by converting the two-dimensional deformation of the azimuth and distance obtained by the radar image into the two-dimensional deformation of the east-west and north-south directions,and comparing with the two-dimensional deformation of the east-west and north-south directions obtained by the optical image,it is found that the deformation results of the optical image and the SAR image are similar,and the deformation of the “11·3” Baige landslide during this period is as high as12.89 m.3.The elevation deformation map after the occurrence of Baige secondary landslide from January 26,2020 to February 4,2022 was obtained by UAV image,and compared with the twodimensional deformation results obtained by COSI-Corr software,the deformation area with obvious slip signs was obtained,and the cumulative deformation of some residual deposits in the east-west direction was up to 7.71 m,which proved the possibility of re-disaster risk and verified the feasibility of optical image offset tracking method to detect surface deformation.4.The COSI-Corr software and CARST software were used to analyze the single deformation of the Slumgullion landslide.The results showed that the east-west deformation of the landslide from 2016 to 2020 was as high as 17.71 m,and the north-south deformation was as high as 16.78 m.The time series displacement of Slumgullion landslide from 2016 to 2020 was obtained by using Landsat-8 image,Sentienl-2 image and the image combined with Landsat-8 and Sentinel-2.By comparing and analyzing the deformation results,it is found that the dual platform not only improves the time resolution of the time series image,but also improves the observation accuracy of the results,which confirms the effectiveness of the dual platform to obtain the time series deformation results of the landslide.