Research On Key Issues Of Landslide Deformation Monitoring And Landslide Mechanism Analysis With InSAR Technique

The landslide failture may cause serious casualties and heavy property losses.The prevention of landslide disaster is an important subject of current researches.The identification and monitoring of landslides are the necessary conditions for the disaster prevention.The Synthetic Aperture Radar Interferometry(InSAR)has a continuous development in recent decades.It has characteristics of wide coverage,high monitoring accuracy,all-weather operations,etc.,and it has been widely applied to the detect,monitor and model many kinds of geological phenomena,such as earthquake,landslide and volcano.Coherence of interferometric phase is a prerequisite for obtaining reliable deformation.In order to reduce the impact of decorrelation and obtain long-term time-series deformation,many multi-temporal InSAR techniques have been proposed(for example,the Permanent-Scatterer InSAR method and Small Baseline Subsets InSAR method).However,landslides usually occur in mountainous areas with large terrain fluctuation,and interferograms are often affected by various types of errors,such as decorrelation noise,atmospheric delay,DEM error,and unwrapping error,which makes it difficult to obtain accurate deformation results by multi-temporal InSAR techniques.Although there are a variety of error correction theories and deformation inversion algorithms,the applications of InSAR in the field of landslide research has been significantly developed,it is difficult to find a method that works well in all situations.Therefore,according to actual situation,designing error correction algorithms and formulating data processing procedures to obtain InSAR landslide monitoring results accurately is an issue that needs to be discussed.In addition,the mechanism interpretation of landslide can deepen the understanding of landslide disasters and provide key information for the prevention of similar cases.Modeling and machine interpretation of landslides through monitoring results is also an important topic that needs to be studied.Based on the above analysis,this dissertation make the research on InSAR landslide deformation monitoring and machine interpretation,discussed the characteristics of several common errors in landslide monitoring,and designed correction methods or deformation processing procedures for these errors.In addition,this dissertation also discussed,analyzed and interpret the mechanism of landslides based on InSAR deformation,precipitation,terrain and geophysical models.This study expands the application of InSAR in landslide mechanism interpretation.The main research contents and results of this dissertation include:(1)The difficulties encountered by InSAR landslide monitoring are systematically analyzed.Several common errors in InSAR landslide monitoring are analyzed,and the basic correction theories of these errors are discussed.The commonly used deformation inversion methods are briefly introduced.In the case of obvious deformation in the north-south direction,based on the data from ascending and descending tracks,the formula for calculating the eastwest deformation is derived.It is proposed that the actual environment and SAR data should be considered comprehensively in the research,and reasonable selection of error correction method and InSAR monitoring algorithm is very important for the data processing.(2)In order to reduce the influence of unwrapping errors on the time-series results,an automatic correction method of unwrapping errors based on Quasi-Accurate Detection(QUAD)is proposed under the framework of SBAS-InSAR.Based on simulation and real data experiments,it can be found that the unwrapping error correction algorithm designed in this dissertation can identify and correct the unwrapping error effectively.(3)The landslide identification and monitoring research is carried out in the area where the interferogram can keep continuous coherence.A landslide identification processing based on SAR Interferometric Point Target Analysis(IPTA)and a SBAS-InSAR landslide monitoring processing based on the optimal interferogram are applied in the study area.A target points selection strategy through multiple threshold is used to select the monitoring points in the IPTA processing.And the suspected landslide regions are mapped by the deformation and the optical image.The method designed in this paper is used to detect landslides in the Wudongde reservoir area.In addition,the type and the driving factor of the Jinpingzi landslide is analyzed and discussed with the spatial and temporal deformation characteristics,precipitation and water level.(4)The landslide monitoring research is carried out in the mountainous area where the interferogram can not keep the continuous coherence.For some area,due to the seasonal precipitation,the interferogram may lose coherence in rainy season.In addition,the huge topographic relief will cause significant tropospheric delay for interferogram.These will seriously reduce the accuracy of the results.To solve the above two issues,this research proposed the concept of effective coherence coverage,and designed an interferogram selection method to filter out interferograms with severe decorrelation;Secondly,an improved quadtree interferogram segmentation processing method is proposed to reduce the influence of atmospheric noise on the deformation monitoring results.It includes the steps of interferogram segmentation,segments reference points selection,deformation inversion and segments merging.Based on the interferogram selection method and segment processing method,the deformation characteristics of the creeping landslides in El Dorado National Forest of the United States are firstly revealed.(5)The mechanism of landslides is further analyzed and discussed based on InSAR monitoring results,external data and geophysical models;Based on the archived SAR data,the historical deformation of the two collapsed landslides are recovered,and the mechanisms of these two landslides are discussed and analyzed through various data;The elastic dislocation model is applied to invert slip distribution of the sliding surface of collapsed landslide.Based on one-dimensional diffusion model and precipitation data,the transient pore water pressure of landslide sliding surface is also modeled.And the driving factors of regional landslide movement are discussed and analyzed with the precipitation,pore water pressure and InSAR time-series deformation.The geophysical model can provide a new insight for the landslide mechanism research,and deepens the understanding of landslide collapsed mechanism.