Surface Deformation Of Xinjiang Axi Mining Area Based On InSAR Monitoring And Analysis

Mineral resources play an extremely important role in promoting economic development,but often cause many geological disasters and safety problems in the mining process,such as destroying surface vegetation,causing land subsidence,destroying the stability of buildings,etc.,and even affecting people’s life and property safety.Therefore,effective,high-precision and multi-dimensional monitoring and prediction of surface deformation is of great significance to grasp the spatio-temporal evolution of surface deformation in mining areas,future development trend and disaster prevention and reduction.In recent years,Interferometry Synthetic Aperture Radar(InSAR)technology has been widely used in many deformation monitoring fields.However,there are many defects in its application,such as spatiotemporal incoherence,phase unwinding error and Digital Elevation Model(DEM)error,which will affect the results of surface deformation monitoring.Therefore,the proposed time-series InSAR technology can effectively overcome the above problems,and also improve the accuracy of InSAR deformation monitoring.However,it is still a problem to be solved how to further combine temporal InSAR technology with mining area deformation and deformation prediction to make it play a real role in providing data support for disaster prevention and reduction in mining areas.Aiming at the above problems,this paper takes Axi mining area in Yining County,Xinjiang Uygur Autonomous Region as the research object,and integrates InSAR time-series technology,Sentinel-1A image data,mining area geological structure and machine learning knowledge to conduct high-precision deformation monitoring,spatial evolution analysis and deformation mechanism analysis of Axi mining area.Furthermore,three different prediction models are used to analyze the deformation prediction.It provides fast,high-precision monitoring method support and reliable data support for the wide application of InSAR technology and disaster prevention and reduction work in mining areas.The main research contents of this paper include:(1)Summarize the principle,research status,error sources and problems existing in application of InSAR technologyBased on Differential Interferometry Synthetic Aperture Radar(D-InSAR)and various time-series InSAR techniques,The current domestic and foreign research status and existing problems of temporal InSAR technology in the field of deformation monitoring are analyzed,and then the principles and error sources of InSAR deformation monitoring are introduced,and the two most widely used temporal InSAR technologies are introduced in detail,as well as the three-dimensional deformation decomposition methods are described.It provides theoretical and methodological reference for the follow-up work.(2)Based on time-series InSAR technology,the temporal and spatial evolution and deformation mechanism of surface deformation characteristics in mining areas are analyzedBased on Small Base Line Subset InSAR(SBAS-InSAR)technology combined with2017-2021 Sentinel-1A high-orbit radar image data,the surface deformation monitoring of Axi mining area in Xinjiang Province is carried out.Then the spatiotemporal evolution and deformation mechanism of the surface deformation are analyzed.The results show that during the whole monitoring period,the surface deformation of Axi mining area is basically stable,and the local surface deformation is more obvious.In addition,mining engineering is the main factor causing surface deformation in mining area,and geological structure,precipitation and water storage in open-pit mine are important factors affecting and intensifying the surface deformation process in mining area.(3)Using multi-track SAR image data to estimate and analyze two-dimensional temporal deformation in Axi mining areaIt is difficult to obtain the complete deformation field of the mining area by using an orbital data,because it can only obtain the one-dimensional deformation of radar LOS,which is difficult to reflect the deformation characteristics of the mining area.Therefore,this paper fuses multi-orbit SAR image data to estimate two-dimensional temporal deformation in the study area.The results show that the average annual vertical deformation rate of Axi mining area in Xinjiang is-10.49 ～ 4.41 mm/a during 2017-2020,and the average annual east-west deformation rate is-4.25 ～ 3.75 mm/a.The average annual deformation rate in the east-west direction of Axi mining area is obviously lower than that in the vertical direction,which indicates that the surface deformation in Axi mining area is mainly in the vertical direction.However,in terms of spatial and temporal distribution,the annual average vertical deformation rate and the east-west deformation rate have high consistency and synchronicity.To sum up,the vertical deformation of Axi mining area is the most obvious,the deformation rate is the largest,the east-west deformation is not obvious,and the north-south deformation is extremely insensitive,so this paper will study the vertical and east-west surface deformation of Axi mining area.(4)Three different prediction models are used to predict and analyze the future trend change of mining area deformationThe monitoring results of surface deformation in Xinjiang Axi mining area based on temporal InSAR technology are combined with the temporal and spatial evolution characteristics of surface deformation in the mining area and the characteristics of time series data set of deformation results obtained by temporal InSAR technology monitoring.Three prediction models,namely grey model(GM),support vector machine regression(SVR)and grey support vector machine regression(GM-SVR),are used to predict and analyze the temporal InSAR deformation monitoring results of Axi mining area in Xinjiang Province from February 21,2017 to February 12,2021.The prediction results show that the prediction effect of GM model is the worst,and the difference between GM model and the real value is large.SVR model is better than grey model,but its accuracy is not high and its stability is low.The prediction effect of GM-SVR model is the best,RMSE is better than 4.02 mm,MAE is less than1.55 mm,indicating that the prediction result is reliable.Therefore,the GM-SVR model can be used to predict the surface deformation of mining area and provide theoretical support for disaster prevention and reduction in mining area.