Research On The Technology Of Signal Processing And Simulation Of Geosynchronous SAR

With the ultra-long synthetic aperture time,ultra-high operating orbit,curve trajectory and the multi working configurations,geosynchronous synthetic aperture radar(GEO SAR)has great advantages in earth's observation,natural disasters' detection,moving target indication(MTI),interferometry,differential interferometry and so on.While,great advantages always come along with great chanlleges.Most classical SAR processing theories are not accurate enough in GEO SAR survey.In this thesis,research on signal simulation and processing of the GEO SAR is carried out.The influences of the characters mentioned above on the GEO SAR exact raw data generation,GEO SAR image formation,tropospheric delay influence on GEO SAR configuration and compensation of tropospheric delay are fully considered.The propose is to realize the high-accuracy simulation of the whole work flow of GEO SAR from raw data generation to the image formation and influence compensation.We hope the theories and conclusions in this article can provide a support to the real GEO SAR missions and further applications.The highlight points and novelties of the thesis are as follows:In Section 3,an accurate raw data simulation algorithm based on the reverse backprojection(ReBP)is proposed.In GEO SAR survey,with the influence of the curve trajectory and earth's rotation,serious range and azimuth variances are introduced into echo raw data,and the simulation of which has become the basic problem of GEO SAR signal processing.Just like its backprojection(BP)sibling,the algorithm is seen as the golden standard for reverse SAR processing due to its exactness in the accommodation of arbitrary surveys,resolutions,swaths,and space-variances.Furthermore,after modifying the beam projection,ReBP algorithm can be widely used to simulate the multi-mode echo data of bi-static,ultra high resolution,medium-high orbit,and GEO SAR configurations.In Section 4,the Range Doppler-Azimuth Chirp Scaling(RD-ACS)and ?K-3ACS algorithms are proposed to compensate the two dimensional spatial variances in GEO SAR survey.Since the GEO SAR raw data is slow varying along range and azimuth directions,the two dimensional spatial-variant high-order Taylor expansion slant range model is built.RD-ACS algorithm is performed in the weak azimuth variance case,with RD compensating the range variance and ACS compensating the azimuth variance,respectively.?K-3ACS algorithm is carried out in the serious azimuth variance case,with?K compensating the range variance and 3ACS compensating the azimuth variance.The focusing image with 2m resolution under 550 s integration time in GEO SAR configuration is achieved with ?K-3ACS.In addition,2D space variant high-order slant range model is introduced into BP flow.By improving the phase calculation,beam projection and interpolation process,we successfully reduce the computational cost and reach a relative high efficiency.In Section 5,considering the long integration time and wide swath of GEO SAR,a tropospheric delay model with variation within both space and time domains is built,and the influence of which on GEO SAR survey is also analysed.The tropospheric delay is divided into the background component and stochastic component.First,we introduce the global pressure and temperature 2 wet(GPT2w)model,based on the Saastamoinen model,Askne model and Vienna mapping function,the background tropospheric delay varying in both the time and space domain is achieved;Second,we introduce the Mat(?)rn function and random walk process to model the space and time variations of the stochastic tropospheric delay.Analysis about the tropospheric delay influence to the GEO SAR raw data and image formation is carried out subsequently with different bandwidth and synthetic aperture time.Finally,simulation feeding the Sentinel-1a image data into the ReBP algorithm with the influence of tropospheric delay based on the proposed model is performed.Comparison between the original image and refocused image validate the correctness of the tropospheric delay model.In Section 6,based on phase gradient algorithm(PGA)and map drift algorithm(MDA),we first achieve the estimation of the two dimensional tropospheric delay phase error.In the first,after divided the image into subswathes and subblocks,modified MDA is performed to reach the whole variant trend of the two-dimensional(2D)phase error.Second,the point-like targets detection is carried out,and the local accurate phases are reached by the modified PGA near the detected targets.Then the least squares fitting and interpolation are used to combine the phase information from MDA and PGA.Finally,phase error compensation is realized by phase multiplication in time and frequency domains.