Investigation And High Accuracy Monitoring Of Landslide Based On High Coherence Points InSAR

InSAR(Synthetic Aperture Radar Interferometry), as a new technology developing rapidly in recent twenty years, has been widely used in monitoring the large scope surface deformation. It has the advantage of all weather, all day, high monitoring precision, large coverage range. Differential Synthetic Aperture Radar Interferometry(D-InSAR) developed on the basis of InSAR, due to its high deformation monitoring precision, high spatial resolution, high operation automation, large monitoring area, little weather effect, has become a very promising technology in the field of monitoring geological disaster. But, due to the temporal and spatial decorrelation and the delay of atmosphere, D-InSAR is seriously restricted the application of monitoring the surface deformation. While, Corner Reflector Interferometry(CR-InSAR) and Permanent Scatterers Interferometry(PS-InSAR) both monitor the deformation of coherence targets to overcome the traditional decorrelation, and make difference of the interferometric phase between coherent points also weaken the influence of atmosphere delay. Therefore, considering the compatibility and complementarity of PS-INSAR and CR-InSAR algorithm, we apply the parameters estimation based on phase coherence, and establish the model of indirect adjustment with constraints to realize the combination of PS-INSAR and CR-InSAR. Then robust estimation is used to weaken the influence of gross errors in observations. In summary, the main researches and innovations are summarized as follows:1. The research of comparison between several PSs identification methods. Common PSs detection methods are: coherence threshold method, amplitude average threshold method and amplitude deviation threshold method. According to the different identification principle, analyze their advantages and disadvantages, in practical application. We need to choose the appropriate method or combine several above methods according to the actual situation of the study area.2. The research of sub-pixel level identification of CR. Rough identification of CR is done based on Range Doppler model(R-D model), the approximate location of the CRs is obtained, then we open a window around the approximate coordinate, the precise location of the peak value of the amplitude can be calculated by multiple interpolation, the position is precise coordinate of CR.3. The research of free connection network of CRs and PSs. The free connection network take the current coherence point as center, connecting all other effective coherence points within the distance threshold, it will not appear long baselines, and the number of baselines increases obviously, the network is more stable.4. The research of parameters estimation based on phase coherence. In this paper, the deformation rate difference and elevation correction difference on the observation baselines are calculated based on phase coherence, Phase coherence can not only reflect the accuracy of parameter estimation, but evaluate the quality of observation baselines, this method can process the differential interferometric phase and do not need phase unwrapping.5. The research of the nonlinear CR-InSAR algorithm based on phase coherence. The method can avoid unwrapping errors caused by large scale deformation gradient and detect large scale deformation, as the distance between CRs is closed, the atmospheric phase effects can be ignored. So nonlinear deformation can be extracted, it reflects the real deformation characteristics of the CRs.6. The research of the combination of PS-INSAR and CR-InSAR algorithm. We take the high accuracy results of CRs as conditions of PSs network to establish the model of indirect adjustment with constraints, taking advantage of the high stability characteristics of CRs itself takes good overall control effect on the solution of whole network to realize the combination of PS-INSAR and CR-In SAR.7. The research of robust estimation applied to weaken the influence of gross errors in observations. This method drops the weight of observation with gross errors, and puts the gross errors into the stochastic model, can not only resist the bias of model, but also can resist gross errors, making the calculation meet the optimization or near optimization in the robust premise, weakening the influence of gross errors in observations.8. The research of applying the combination of CR-InSAR and PS-InSAR algorithm to the Jiaju landslide deformation monitoring. The deformation rate results of Jiaju landslide are obtained based on this method, and compare to the existing survey data, the deformation results match well with the landslide zoning characteristics, it not only proves the d reliability of the algorithm, but also reveals that Jiaju landslide has obvious deformation characteristics reasonably.