Research On Early Identification And Surface Deformation Of Potential Landslides From Diexi To Feihong Section Along The Upper Reaches Of Minjiang River Based On InSAR Technology

Mao County is settled in the mountainous area,southwest China,northwest of Sichuan Province.The study area has steep terrain,complex geological structure,more common and various types of geological disasters.In the past decade,the occurrence frequency of landslides has accounted for about 71.6%of the total number of disasters in China,with the characteristics of suddenness,strong destructiveness and great harm.There have been many landslide incidents in Mao County.These disasters not only disrupted road traffic and caused huge property losses,but also pose a threat to human life and safety.For example,the Xinmo Village landslide in 2017caused river blockage,buried people and huge economic losses.Therefore,the effective identification of potential landslides in the study area is of great significance to the safety of residents'lives and property and disaster prevention and mitigation.In this paper,we take the section from Diexi to Feihong in Mao County as the research object.The 38 Sentinel-1A ascending data covering the study area were used as the data source.In order to obtain the surface deformation,we used the SBAS-In SAR,which has high-precision,to monitor the study area.The annual average surface deformation rate in the period is obtained.Due to the side-view imaging system of the satellite,SAR images will have geometric distortions.Therefore,the paper analyzed the visibility of SAR images in the study area based on DEM and satellite parameters,and retained the In SAR monitoring data in the visibility area.Then,using hotspot and kernel density analysis to perform spatial statistical analysis on the reliable deformation points in the visibility area,the potential landslide area in the study area is obtained.Finally,according to the deformation rate obtained by the time-series In SAR technology,optical remote sensing images and field investigation verification,we determined the potential landslide points in the study area.The main achievements are as follows:(1)According to the topography and geomorphology and geological conditions of the study area,we used the SBAS-In SAR technology and selected appropriate parameters to process the Sentinel-1A ascending data.Then,we obtained the annual average deformation rate of the study area from 14 October,2014 to 13 August 2019along the radar line of sight(LOS)to the surface.It can be seen that the whole research area is basically stable,and the deformation only occurs in a small range,especially along the banks of the Minjiang River.The rate is mainly between-46.50mm/y and 39.87mm/y,mostly concentrated in-10mm/y?10mm/y.(2)The slope and aspect of the DEM of the study area were analyzed.Combined with the incident angle and flight mode of the satellite,and used Arc GIS software to divide the study area into five areas:good visible area,low sensitive area,foreshortening,layover and shadow.Considering the reliability of the monitoring results,the good visible area and the low sensitive area are divided into visible areas,and other areas are divided into invisible areas.Among them,the area of the visible area is 377.7km~2,and the invisible area is 234.1km~2,and the invisible area is mostly on the slope facing southwest.The results proved the superiority of Sentinel-1A data in the study area.(3)Using the spatial analysis function in Arc GIS software,the hot spot analysis of the deformation rate in the visible area was carried out to obtain the aggregation degree of the deformation points and eliminated the points with no statistical significance.Then,the kernel density analysis is carried out on the basis of it,highlighting the abnormal deformation area in the study area and visualizing it.A total of 18 potential landslide areas were acquired,which were used for subsequent identification of potential landslides,most of which were distributed on both banks of the Minjiang River.(4)We summarized signs of landslide identification.Through images,SBAS-In SAR deformation rate,field investgation verification using UAV and field surveys,etc.,a total of 20 potential landslides were finally identified.Among them,there are 7 old landslide points and 13 new landslide points.We counted the slope,aspect and elevation of each landslide and found that the slope of potential landslide is mostly about 30°,and the height difference is large.The research shows that the combination of SBAS-In SAR technology and multiple spatial analysis can effectively monitor and accurately identify reep-type potential landslides,and provide the basis for the government's disaster prevention and mitigation work.