| Urban surface deformation is an environmental geological problem which is ubiquitous in the process of urbanization in today's world.This continuous,gradual and cumulative geological disaster,whose process is slow and irreversible,has become an important factor restricting the sustainable development of regional economy and affecting the safety of people's life and property.DInSAR,as an effective surface deformation monitoring technique for large area with high spatial resolution,is an effective supplement to traditional ground measurement technology(such as leveling,total station and GPS)which is local measurement by points.It has advantages of all-day,all-weather,wide range and high precision.However,the uninsurable problem of InSAR technology is the ability to obtain only one-dimensional surface of information that can be accessed by radar,unable to detect the actual 3D deformation of the surface.In reality,not just vertical deformation affects the security of the city,but horizontal stretching or extrusion can pose a threat to pipelines,railways,metro and buildings.Therefore,It is one of the important contents of constructing scientific and efficient urban ground subsidence monitoring system to obtain urban 2D or 3D deformation information.Based on high resolution SAR images,2D time-series deformation monitoring of urban surface has advantages of short revisiting period,high spatial sampling and high precision positioning.In addition,the multidimensional small baseline subset deformation monitoring technology(MSBAS)is developed on the research of 3D deformation field or 3D deformation rate.The model is simple,and the precision of the solution is irrelevant to the time of image acquisition,and it has been applied to the measurement of coal mining,urban development,carbon storage,permafrost evolution and volcanic activity.However,when multi-track fusion is adopted,the spatial sampling density is improved by the interpolation method of the interferogram,and interpolation error is introduced,which affects the accuracy of the algorithm.Without interpolation,the temporal sampling(interferograms selection)and spatial sampling(like the pixelmask)restrict each other,which reduces the integrity of the deformation information extraction to a certain extent.Besides,in the practical deformation monitoring application using high resolution image,the difference between the real elevation of the ground objects and the DEM leads to the discontinuity of the differential interferogram phase between the adjacent pixels,which is difficult to satisfy the unwrapping conditions and affects the 2D deformation monitoring results.The current research methods are generally to analyze topographic error correction for medium resolution images or local small regions,and the research on correction of terrain errors in urban areas with high resolution SAR images is relatively deficient.Moreover,in the calculation of topographic error,the researchers focused on the data source to estimate the topographic error.At present,the model is not able to balance the complexity and the precision of the model in solving the error correction whether using the wrapping data(differential interferometric phase)or unwrapped data(the unwrapped phase or the deformation rate).In view of the above problems,this article studies the topographic error correction model and 2D time-series deformation model,puts forward the solution,and combines with actual application case to verify feasibility of the model.The specific research contents are as follows:(1)From the perspective of InSAR,the development of spaceborne SAR systems are summarized,which mainly involves the technique principles,the development,the key parameters and the mission objectives of SAR system.The paper combs the most commonly used time-series analysis technology and analyzes the applicability and limitations of different technical methods and the deformation features of the application target by some applications.Systematic analysis of SAR system,technical method and application condition provide theoretical support for subsequent research.(2)Based on the analysis of the effects of topographic error on image registration,phase unwrapping and geocoding,this paper summarizes the research status of topographic error correction in InSAR time-series analysis and improves the limitations of existed techniques in the application of high resolution image InSAR timing analysis.Firstly,the influence of topographic error in the high resolution InSAR time-series deformation technique is impaired using the method of variable spatial baseline and temporal baseline,which is based on the InSAR time-series algorithm model and the relationship between the spatial baseline and the height ambiguity.Then,the model parameters are optimized according to the average elevation and deformation characteristics of Shanghai high-rise buildings and urban subsidence area.It is proved that that topography error in the model is gradually corrected by the unwrapped interferograms and the height residual error maps,and the feasibility and accuracy of the method are verified for the high building district and the subsidence area in Shanghai by comparison of the height error histograms.(3)The paper summarizes the problems of the multi-track fusion 2D time-series deformation models.According to the research purpose,the key technology in MSBAS model,including the adjustment of different tracks,observations and model parameter optimization,interferograms selection,are improved to realize temporal sampling,spatial sampling and calculation accuracy in the multi-track 2D time-series model.Firstly,for the reference correction between multiple tracks,the ground subsidence rate map and the corresponding histograms in the central area of Shanghai obtained by the ascending and the descending data sets show the difference when reference standards are not consistent in different tracks,and the statistics method for reference calibration is proposed.The observations and the incidence angle of the model is the precondition of high precision calculating.The observations of the 2D time-series deformation model in this paper are adopted from the differential interferograms after topographic error correction,and the incidence is optimized to the 2D incidence corresponding to the image pixel.In addition,in order to solve the problem that the temporal sampling and spatial sampling of MSBAS can not be taken into account simultaneously,we propose a method of establishing the observation matrix for each pixel and then eliminating the interferograms with low coherence,to improve the spatial and temporal sampling ratio of the data sets.(4)In Shanghai urban area,the two groups of interferogram sets with multi-looked spatial resolution of 5 m and 15 m are used to experiment,and the deformation rate maps in the east-west and up-down directions,and 2D time-series deformation line graph of multiple pixel positions are obtained.Compared with the results of the original model,it is found that the ratio of pixels participated in calculation increases from 30% to 75%.In addition,the deformation results of this paper are compared and analyzed with the data of ground subsidence monitoring station.The deformation rate in the east-west direction is basically consistent with the GPS monitoring results,and the small difference may result from the inconsistency of the observation interval and the influence of the nonlinear deformation characteristics during long time periods.Compared with the ground subsidence monitoring station,the vertical accumulation deformation has a good trend consistency,which verifies the reliability of the method used in urban 2D temporal deformation monitoring.(5)The study is on the application of 2D time series deformation monitoring model.Taking the section of subway line 8 and line 10 in Shanghai as the research object,the article summarizes the deformation constraints of the engineering construction,the regional strata distribution,the engineering geology and the hydrogeology characteristics firstly.Combined with linear engineering and their 2D time-series deformation monitoring results,we analyse the dynamic characteristics and the basic regulations of deformation development.It is found that the deformation velocity in the east-west direction along the profile veries gradual,and the accumulated deformation varies from two to five millimetres temporally.The spatial distribution of the deformation rate in the up-down direction is inhomogeneous,and the deformation rate along the subway line is greater than that of both sides.The deformation of different feature points has the consistency characteristics in temporal distribution. |
PDF Full Text Request