Research On Data Fusion Method In Ground-based SAR Landslide Monitoring

Ground-based synthetic aperture radar is a new microwave remote sensing technology which developed in the past ten years.It has strong penetrating power and can monitor continuously under bad weather.It has sub-millimeter precision in surface deformation measurement,and has strong advantage in the field of landslide monitoring.But the reliability of deformation may decline for the phase instability.In this paper,permanent scattering point selection can select monitoring points with high quality phase,which aim to ensure the reliability of deformation.The data fusion is a technology which use a variety of data information to fuse through a mathematical integration model.This technology is originated in the 1970 s,whose application field is extended from military to civilian gradually.In this paper,using the complementary advantages of Global Navigation Satellite System deformation data,optical image and Digital Elevation Model,so that obtain multi-level and multi-angle information,provide a simpler and more direct help,and achieve the purpose of effectively assisting in early disaster warning.The main works of this paper is as follows:(1)Permanent Scattering(PS)selection of double threshold Based on the average correlation coefficient and the amplitude dispersionFor the problem that PS's number and quality can't be guarantee at the same time in traditional methods.In this paper,we adopt dual-threshold method to select PSs and extract process parameters.Firstly,using the average correlation coefficient method to obtain Permanent Scattering Candidate(PSC)points.The adaptive threshold is adopted in this step,it can save the number of PSC points.Secandly,using the amplitude deviation method to select PS points in PSC points.The optimal parameter value is adopted in this secondary step.Finally,obtaining the PS points with better quality and number.The experimental results show that our method of PS points can effectively eliminate targets which has poor phase quality,serious miscorrelation in this paper.It can improve the phase stability of the deformation calculation target,and ensure the accuracy of the ground-based SAR deformation measurement results accuracy.(2)Deformation fusion method of Ground-based Synthetic Aperture Radar(SAR)and Global Navigation Satellite System(GNSS)based on Newton's methodFor the disadvantage of single interpretation of ground-based SAR deformation data.In this paper,we applies the fusion method of In SAR and Global Positioning System(GPS)deformation data to the ground-based SAR and GNSS.Firstly,determining the position of the GNSS monitoring point in the two-dimensional grid of ground-based SAR.Secondly,performing Kriging interpolation on the GNSS three-dimensional deformation,and obtaining the three-dimensional interpolation deformation of the ground-based SAR.Then,analyzing the geometric relationship between the ground-based SAR's line-of-sight direction deformation and GNSS's the three-dimensional terrain,constructing the energy function which regard the ground-based SAR's real three-dimensional deformation as unknown quantity.Finally,calculating the ground-based SAR real terrain directional deformation value by the direct analysis method and the Newton method.Errors are evaluated for the quality of deformation fusion.The experimental results show that the accuracy and reliability of our method is better than the interpolation and direct analysis methods,which improves the application ability of ground-based SAR from onedimensional deformation to three-dimensional deformation effectively.(3)Image fusion of Ground-based SAR image and optical image based on Digital Elevation Model(DEM)For the shortcomings that ground-based SAR images' scattering targets of ground objects is difficult to identify.In this paper,a data fusion method of ground-based SAR image and optical image based on digital elevation model is proposed,and the coordinate transformation of different data sources is performed uniformly.Firstly,using the fourparameter method to transform the two-dimensional coordinates of the ground-based SAR image,and using the digital elevation model and its matching mapping to obtain the threedimensional ground-based SAR image.Secondly,using the pixel scaling transformation to perform the two-dimensional coordinate transformation of the optical image,and then using the digital elevation model perform regional iterative projection to obtain three-dimensional optical images.Finally,the ground-based SAR image and deformation data are matched and mapped to the corresponding optical images on the basis of the coordinate unification of each data.And then the fusion result information is obtained.The experimental results show that the method in this paper can combine the color,deformation monitoring ability and resolution complementarity of the two images effectively.It can realize the target recognition of the ground-based SAR monitoring area.Thereby,enhancing the external application capability of ground-based SAR monitoring.