Surface Subsidence Monitoring And Driving Force Analysis Along The Nanchang Subway Line Based On InSAR Technology

Due to the traffic pressure and the demand of population movement,the construction of subway has become an important goal of urban construction.However,in the construction and operation of subways,surface settlement is inevitable.If the settlement exceeds the safety limits of deformation,it will bring incalculable harm to people’s personal safety and economic development.Synthetic aperture radar interferometry(In SAR),with its large monitoring range,high spatial and temporal resolution,and non-contact characteristics,has been widely used in monitoring geological hazards such as earthquakes,landslides,and surface subsidence.In this paper,we use PS-In SAR technology and SBAS-In SAR technology to monitor and extract the subsidence along the subway in Nanchang area,and use the geodetector model to perform the discretization and factor detection for five factors:road density,water density,elevation,soil and lithology,factor interaction detection,and construct the ground subsidence sensitivity map based on the convolutional neural network U-net model.The main research is as follows.(1)The 68-view Sentinel-1A uplift image data of Nanchang main urban area were processed by SBAS-In SAR technique and PS-In SAR technique to obtain the subsidence rate and cumulative deformation from June 2017 to December 2021,and 34 pairs of image interference pairs were selected to calculate the mean value of coherence coefficient,which were all greater than 0.5.The SBAS-In SAR and PS-In SAR results were extracted.In SAR and PS-In SAR results,30 random common points were extracted,and the deformation trends tended to be consistent,and the correlation coefficient was 0.84563 by Pearson correlation analysis,which verified the reliability of the subsidence information obtained by SBAS-In SAR.(2)Based on SBAS-In SAR technology,the deformation maps within the 2-km buffer zone along the four metro lines in the study area were extracted.Among them,a large subsidence funnel area exists on both sides of the road along the interval from Kongmu Lake to Changjiang Road Station of Line 1,and the deformation trend is more obvious.The settlement area of Line 2 is distributed in South Road Station and Nanchang East Station,and the southwest of Line 2 is Nanchang Jiulong Lake Development Zone,where Jiulong Lake is in the development stage and the settlement may be related to the construction.Line3 settlement area is mainly distributed in the middle section and the southern section,the middle section of Jiangling station in April 2020 showed a strong settlement trend,the southern section of the Silver Triangle North station settlement trend is more moderate.The settlement area of Line 4 is distributed between Yufeng Street and Lizhuang Mountain area,and between Qili Street and Fishtail Island area,where the settlement phenomenon is more serious and dangerous,and the accumulated deformation all reach-65 mm.(3)Road density,water density,soil,lithology and elevation were selected as the driving factors,and the correlation between the spatial distribution of each driving factor and the subsidence rate was explained by using the divergence of geographic detectors and factor detection and interaction detection.The subsidence sensitivity map was drawn using U-net model prediction,and the results showed that the area of very high sensitive area of Line 1accounted for 45.48%,the area of very high sensitive area of Line 2 was the highest,the area of low sensitive area of Metro Line 3 was the highest distribution,and the area of very high sensitive area of Line 4 was the least.