Study On Landslide Warning Method And Model Based On GB-DInSAR

Landslide is a kind of common natural geological disaster,which has caused great loss to the regional economic construction and people's life all over the world.Therefore,the monitoring,forecasting and early prevention and control of landslides have been widely concerned and studied all over the world.In recent years,ground-based DIn SAR,as a non-contact high-precision monitoring means,has been widely concerned and applied with its wide range,high accuracy,all-day,etc.However,there are relatively few researches on landslide monitoring and warning methods based on DIn SAR,so it is urgent to carry out relevant researches.In this paper,based on the conventional point-based monitoring model and method for landslide monitoring and early warning,combined with the characteristics of ground-based DIn SAR areal monitoring and data diversification,the landslide prediction model and method suitable for it are deeply studied and explored.The research is mainly carried out from three aspects: baseline error correction,landslide identification and prediction,and landslide time prediction.The main contents include:1.The basic theory of ground DIn SAR slope deformation monitoring is briefly introduced.The factors affecting the monitoring accuracy are mainly studied from the aspects of imaging observation geometry and decoherence.2.To solve the problem that the main slide direction of regional slope landslide is difficult to predict due to the limited monitoring dimension of DIn SAR apparent deformation of conventional foundation,a two-field gradient method is proposed to identify and predict the main slide direction of slope landslide.The deformation field and terrain field in the landslide area are analyzed,the surface fitting method based on local translation is developed,the analytical expressions of the deformation field and terrain field in the monitoring area are derived,the identification and prediction of the main slide direction are realized by the gradient change of the deformation field and terrain field,and the validity of the method is verified by the actual data.3.Aiming at the problem of existing ground-based DIn SAR landslide prediction model is single,which mainly uses the threshold detection method to provide landslide prompts and other shortcomings,a landslide early-warning model is established to reflect the stability of the landslide body based on three parameters--extent of change(EOC),extent of stability(EOS),and extent of mutation(EOM).And the nonlinear inverse relationship between the target deformation velocity and coherence is obtained.On this basis,a landslide time prediction method based on coherence change detection is proposedand applied to the DIn SAR system of actual foundation.Finally,the accuracy and effectiveness of the model and method in DIn SAR monitoring and warning on specific ground are verified through actual monitoring data,which provides accurate and timely warning information for decision makers.