Surface Deformation Monitoring And Soil Salinization Inversion Based On SBAS-InSAR In Saline Soil Area

The unfavorable engineering geological characteristics of saline soil,such as salt swelling,dissolution and corrosion,will cause destructive disasters to natural environment and artificial buildings,such as slope instability,building rupture,uneven uplift or subsidence of roadbed and corrosion of underground metal pipes.It is of great significance to carry out high-precision time-series deformation monitoring of saline land surface and inversion of soil salinization for prevention and control of potential geological disasters,rational development of natural environment and guarantee of large-scale engineering construction.However,the traditional monitoring method of surface deformation is to use GPS,total station and other artificial control points,which is time-consuming and laborious and difficult to monitor large area long time series deformation.The lack of high precision deformation information of large range long time series and the complexity of deformation in saline soil area make it difficult to reveal the overall deformation law of saline soil.At the same time,the current soil salinization inversion models mostly focus on the exponential modeling of multispectral information,but due to the comprehensive characteristics of soil spectral response and the existence of "foreign bodies in the same spectrum" phenomenon,its accuracy is difficult to guarantee.At present,there is still a blank method for evaluating soil salinization degree with deformation information.If the inversion of soil salinization degree can be completed with deformation information,it may make up the limitation of multispectral salinization evaluation index and enrich the theoretical model of soil salinization degree inversion.SBAS-InSAR(Small Baseline Subset Interferometric Synthetic Aperture Radar)is an excellent time-series InSAR technology,which can obtain a wide range of time-series deformation.It has incomparable advantages in monitoring large-scale surface time-series deformation with high precision and high temporal and spatial resolution,and provides an unprecedented opportunity to detect large-scale high-precision time-series deformation of saline land surface.However,due to the characteristic of seasonal deformation in saline soil area,the SBAS-InSAR method is not applicable to the linear modeling of deformation phase in the calculation process,which seriously affects the accuracy and reliability of deformation extraction.Aiming at the above problems and challenges,this paper selects Dunhuang region of Gansu Province as the research object.In-depth research was carried out from four aspects:improvement of SBAS-InSAR model in salinized land area,identification of seasonal deformation rules and causes in salinized land area,construction of soil salinization evaluation index based on deformation information,selection of optimal parameters of soil salinization inversion model based on deformation index.Aiming at the problem that the traditional time-series InSAR method for deformation phase linear modeling is not suitable for the salt soil area with seasonal deformation characteristics,which leads to the low accuracy and reliability of deformation extraction,the SBAS-InSAR algorithm is improved.The traditional linear model of deformation phase in sbas-insar calculation was replaced by trigonometric model and polynomial model with better interpretation of periodic deformation,which refined the separation process of low-frequency deformation phase in saline soil area and achieved the purpose of extracting high spatiotemporal resolution deformation in saline soil area.In order to solve the problem that there are different deformation trends in regions with different soil salinization levels and it is difficult to reveal the overall deformation law,the high-precision time-series deformation information of Dunhuang,a typical saline soil region,was obtained by using 57 C-band Sentinel-1A image data from January 2018 to December2019 based on the improved SBAS-InSAR method.Through the time-series analysis of the characteristic points of different soil salinization degree regions,the internal law of different soil salinization degree regions and seasonal deformation is revealed,which is of great significance to deepen the scientific judgment of engineering characteristics of salinization soil and the prevention and control of potential geological disasters.In view of the problems of unsatisfactory effect of existing remote sensing methods on salinization evaluation and lack of salinization evaluation methods based on deformation information,in this paper,a soil salinization evaluation method,Seasonal deformation salinization Index(SDSI),was established based on the seasonal deformation characteristics of salinization soil.In order to complete the logical closed loop,the correlation analysis and curve fitting analysis based on the principle of statistics were used to compare SDSI and classical multispectral salinization evaluation indexes.The results showed that SDSI’s evaluation of salinization degree was significantly better than other multispectral salinization evaluation indexes,which verified the correctness of SDSI’s theoretical method.The method of soil salinization evaluation based on deformation information was filled in.However,statistical analysis alone is not enough to explain the actual value of a method.It is necessary to use SDSI to invert the saline soil area and draw a convincing conclusion by comparing it with the real surface soil salinization degree.For SDSI inversion salinization degree how fair selection of parameters in the process of problem,the optimal parameters by using accurate traversal easy to overlook the small categories,this article uses enough punishment model "bias" of Kappa coefficient of we instead of accuracy for predicting model,through two traversal and four significant figures of the inversion model of the optimal parameters,finally,SDSI soil salinization degree inversion model was used to complete the inversion of salinization degree in Dunhuang area.The Kappa coefficient of inversion result was 0.3632,and the classification accuracy was 64.75%.At the same time,by comparing the inversion results based on the selection of accuracy parameter,it is found that the selection principle of Kappa coefficient can ensure high accuracy without ignoring small categories.Thus,the feasibility and reliability of the soil salinization evaluation index SDSI based on deformation information are completely proved,and the theoretical model of soil salinization inversion is enriched.