| Landslide is a very common and dangerous geological disaster,which is more commonly distributed in the mountainous areas of southwest China.In recent years,the number of landslide disasters accounts for more than 70% of all geological disasters in China.Among them,there are some mega landslides,which pose a serious threat to the safety of local people’s life and property,as well as the surrounding houses,roads,bridges and other buildings.Therefore,landslide detection and monitoring are of great significance in terms of disaster prevention and emergency response.Traditional means of landslide monitoring include level measurement,GPS observation and remote sensing interpretation based on optical images.However,these methods are more or less limited by operational cost,spatial resolution and weather,and cannot meet the needs of large scale landslide hazard point detection.The emergence of synthetic aperture radar interferometry technology has greatly solved the above problems.It has the characteristics of all-day,all-weather,high spatial resolution and large working range.It is able to penetrate clouds and actively emit and receive microwaves,which has more obvious advantages in fields such as deformation monitoring and landslides.However,the LOS-oriented deformation observation results obtained only from a single platform/orbit cannot effectively reflect the real deformation of the ground surface.In particular,landslides generally move downward in the slope direction,and the direction of their movement may differ greatly from the LOS direction.In order to solve this problem,this paper proposes a slope-oriented deformation solution model with joint ascending/descending track time-series In SAR to obtain the sliding surface deformation field,through which the deformation field can be used to understand more clearly the deformation trend and motion law of landslide body.This paper selects the long strip area along the Jinsha River from Lomai Township,Gongjue County,Tibet to Zhubalong Township,Batang County,Sichuan as the study area to carry out landslide detection and long time series deformation monitoring of the upper reaches of the Jinsha River in Gongjue Section.The specific work is as follows:(1)Landslide hazard point detection in the upper Gomjue section of Jinsha River based on Stacking-In SAR technologySentinel-1A ascending/descending track SAR data spanning the period from January2019 to May 2021 were used to obtain the annual average deformation rate map of LOSoriented ascending/descending tracks in the study area based on the Stacking-In SAR technique.Based on the available data,±10 mm/yr was selected as the surface deformation rate threshold,and the points exceeding this range were regarded as landslide potential points.Then,combining the visual interpretation of Google earth optical images and existing research data,we determine whether the point is a landslide.Finally,the landslide points identified based on the ascending/descending track SAR data are catalogued.A total of 49 landslide potential points were identified using the ascending/descending track data.Among them,33 were identified based on the ascending track data and 38 based on the descending track data;22 landslide potential points were identified based on the ascending/descending track together.There are 11 landslide hazards identified based on the ascending track data alone and 16 landslide hazards identified based on the descending track data alone.(2)Long time series monitoring of landslide complex in Shadong Township area of Jinsha River based on SBAS-In SAR technologyThe landslide monitoring based on SBAS-In SAR technology was carried out for the identified landslides of Xiongba and Sela in the Shadong Township landslide cluster area,where deformation is concentrated.By obtaining the annual average deformation velocity field and the cumulative deformation field of the ascending/descending LOS direction of the two landslides,the regional distribution of the deformation of the landslides was analyzed.And the cumulative subsidence maps of the landslide’s deformation characteristic points were plotted,and it was found that the accelerated deformation period was very obviously related to the rainy season.(3)Landslide body along-slope deformation modeling with joint ascending/descending track time series In SARIn the joint solution of the ascending/descending In SAR deformation observations,the influence of the inconsistent deformation starting bases due to the different reference points selected in the phase unwrapping is eliminated by introducing the systematic deviation K.The geometric relationship between the LOS directional deformation component and the N.E.H.3D deformation component is derived first,and then the geometric relationship between the N.E.H.3D deformation component and the slope direction,vertical slope direction and normal direction is derived.Assuming that the entire sliding surface is uniformly distributed(the default normal deformation is 0),the vertical deformation velocity field with the highest accuracy and the system deviation K are solved first,and then K is substituted into the solution of the slope direction and vertical slope direction deformation velocity field.Finally,the total deformation velocity field on the sliding surface is obtained by vector superposition.(4)Extraction of sliding surface deformation field of Xiongba/Sela landslide based on joint ascending/descending time-series In SAR deformation observationTaking the Xiongba landslide and the Sela landslide as examples,we extracted the sliding surface deformation models of the landslides based on the joint ascending/descending timeseries In SAR,and obtained the vertical,along-slope,vertical-slope and sliding surface deformation velocity fields of the landslides.Based on the results of the model solution,the Xiongba/Sela landslide is analyzed in terms of fault,lithology,geological structure,rainfall and river erosion.Finally,the results of the existing literature are compared with the results of this paper,and the high agreement between the two results indicates that the results of the model are reliable. |
