Early Identification Of Landslides In The Upper Reaches Of Minjiang River Based On InSAR Technology

In the western mountainous area of China,the topography is complex,the geological structure movement is active,and the geological disasters occur frequently,which pose a serious threat to the safety of people's lives and property.In recent years,many large earthquakes have been recorded in the upper reaches of the Minjiang River and its surrounding areas.Large earthquakes have seismic cracking and loosening effect on the mountains,causing a large number of mountain cracks,which worsens the situation of many potential landslides.Therefore,it is urgent to carry out large-scale early identification of potential landslides in the upper reaches of the Minjiang River.InSAR technology has the characteristics of strong penetration,large range and high accuracy.In this paper,sentinel-1A image is taken as the data source,and the upper reaches of Minjiang River is taken as the research area;The terrain visibility of sentinel-1a data in the study area is studied;It enriches the identification rules of potential landslides and carries out the early identification research of landslides;Based on the results of landslide identification,the risk level is divided.The main results and conclusions are as follows:(1)In order to identify the potential landslides in the early stage,a technical method of fusing multi-source data such as SAR image,DEM data and high-resolution optical remote sensing image is established.The terrain visibility of sentinel-1A data in the study area is studied.The results show that the total area of overlay and shadow of sentinel-1A data in the upper reaches of Minjiang River accounts for about 5.14% of the total area of the basin,and about 95.5% of the geological disaster points are in the visible area,which indicates that sentinel-1A data can meet the needs of Surface Deformation Inversion and early landslide identification.(2)Using PS-InSAR and SBAS-InSAR to obtain the results of surface deformation in the upper reaches of Minjiang River,the overall trend of deformation obtained by the two technologies is basically the same,which indicates the effectiveness of InSAR Technology in the monitoring of surface deformation in this region.Based on the analysis of the deformation results,it is found that the major deformation areas are mainly concentrated on both sides of the river and the iceberg area in the high-altitude area.The area near the river valley mainly shows subsidence,and the area far away from the river valley mainly shows the overall characteristics of uplift,which is the result of river erosion,structure and other factors.At the same time,heavy rainfall and earthquake will also cause surface mutation.From the deformation results of the study area,it is found that SBAS is more suitable for the inversion of the area where the target is planar.(3)Based on setting the deformation threshold of unstable slope,the distribution characteristics of disaster points,the hidden danger area,the optical remote sensing characteristics of potential landslides,and the introduction of landslide unit as the reference boundary of potential landslides,the recognition rules enrich the previous recognition rules,which are more reasonable and reliable.According to the identification rules,a total of 283 potential landslides were identified,including 82 new ones.Landslides are mainly concentrated in NDVI < 0.3,altitude < 2500 m and slope <35 ° By superimposing sentinel-2 image,it is found that most of the potential landslides are in cultivated land,winding mountain highway,bare land and other areas.(4)Based on the early identification results of potential landslide,the landslide area and the maximum deformation rate in the area,slope,NDVI and other eight factors,the risk level of potential landslide is divided by using the analytic hierarchy process.The results show that there are 53,120,98 and 12 extremely low risk landslides,low risk landslides,medium risk landslides and high risk landslides respectively,and the area accounts for 8.47%,32.48%,53.59% and 5.46% of all potential landslides respectively,which indicates that the average area of extremely low risk landslides is small,while that of medium risk landslides is large.