Research For InSAR Phase Unwrapping Algorithm Based On Nonlinear Filtering And Wavelet Transform

In SAR calculates the digital elevation model(DEM)of the target area by analyzing the interference phase difference of the SAR image pair in the same geographical area,which is an efficient means for surface deformation investigation and monitoring.In SAR technology has very high accuracy of ground survey,but also is easily affected by various interference sources of the system itself.The result of phase unwrapping is also one of the main error sources in In SAR technology,and it directly affects the accuracy of elevation data acquired afterwards.This paper research on the phase unwrapping problem of interferograms in In SAR system,the research work can be divided into three aspects:1.Research on the traditional phase unwrapping algorithm.The phase unwrapping algorithms of path-following,minimum-norm and nonlinear filtering are analyzed in detail,and the performance of each algorithm is compared between simulation data and In SAR interferogram processing.2.The phase unwrapping algorithm based on the recursive estimation of the local polynomial of the cubature Kalman and the phase unwrapping algorithm based on the local iterative estimation of the cubature information filter are proposed.The gradient parameters in the polynomial coefficient vector are obtained by AMPM gradient estimator,and the state vector is estimated in the local window by cubature Kalman filter,which can effectively process the interferogram with low SNR.The phase unwrapping algorithm based on the local iterative estimation of cubature information filtering improves the efficiency of phase unwrapping by reducing the dimension of state vector and using the cubature information filtering to estimate the phase value.3.A wavelet phase unwrapping method based on cubature Kalman filter is proposed.Wavelet transform is used to decompose the interferogram,and the phase unwrapping algorithm based on the local polynomial recursive estimation of cubature Kalman is used to solve the low-pass coefficient after the phase unwrapping decomposition.The multi-scale analysis of the phase unwrapping problem improves the noise suppression ability of the phase unwrapping of large scale interferogram.