Research On Multi-channel InSAR Elevation Retrieval Based On Information Fusion

Multi-channel Interferometry Synthetic Aperture Radar(InSAR)elevation retrieval technology is a three-dimensional imaging technology developed by breaking through the limitations of traditional single-channel InSAR technology and has attracted more and more scholars to study this.It can obtain high-precision in terrain with discontinuous phases compared with traditional single-channel InSAR technology.This paper mainly studies the multi-channel InSAR elevation retrieval technology based on information fusion.The main research contents are as follows:(1)Researched and analyzed the multi-channel InSAR Chinese Remainder Theorem(CRT)algorithm,the maximum a posteriori(MAP)algorithm,the maximum likelihood(ML)algorithm and related improved algorithms,and many commonly used multi-channel InSAR elevation inversion algorithms.(2)An InSAR elevation inversion algorithm combining ML algorithm and unscented Kalman filter(UKF)algorithm is proposed,incorporating ML estimated elevation information and local phase gradient based on the modified matrix beam model(AMPM)estimate the information by constructing a robust UKF elevation reconstruction system model,thereby breaking through the phase discontinuity condition limit,and the experimental results of elevation reconstruction of different types of terrain verify the effectiveness of the method.(3)A multi-channel UKF elevation inversion method based on phase discontinuity detection strategy,quality map guidance method,AMPM local phase gradient estimation technology and ML algorithm is proposed.This method uses phase discontinuity detection strategy detect discontinuous boundary points,perform multi-channel UKF elevation inversion under the path tracking strategy guided by the quality map,and obtain the elevation estimate of the terrain.It has the preponderances of single-channel and multi-channel InSAR technology at the same time,so that high-precision elevation results can be reversed in continuous and discontinuous terrain,and the effectiveness of the algorithm is proved under multiple experimental terrains.