Study On Imaging Algorithms For Spaceborne LEO/GEO SAR

With the increasing application requirement on imaging performance,the research on spaceborne synthetic aperture radar(SAR)is advanced to high resolution,wide swath and various working mode.In the case of high resolution or the high orbit,satellite will moves along three-dimension(3-D)spatial curved line with the effects of Earth rotation,which results in the conventional imaging model and focusing method are invalid.We start our study as the reconstructing of spaceborne SAR geometry.Then,the problems of low earth orbit SAR(LEO-SAR)and geosynchronous SAR(GEO-SAR),such as the range-azimuth coupling,spectrum aliasing and space variance,etc.are mainly discussed.The main content of this dissertation is summarized as follows.1.In the case of high resolution and large-scene,the trajectory of satellite is not a straight line but a 3-D spatial curved line.Thus,the spaceborne SAR model is invalid on the basis of assumption of rectilinear motion with constant velocity.To solve this problem,an accurate geometry for spaceborne SAR is constructed based on vector analysis and coordinate transformation,and the factors are analyzed deeply,which affect the accuracy of geometry.Then,the expressions of Doppler parameters and resolution are derived,which are important for the design of system parameters and imaging method.In addition,with the increasing requirement on high resolution imaging,the error caused by "stop-go" model cannot be neglected.According to the geometry of transmit-receive,the expression of real propagation delay is derived,and an equivalent midpoint range model is proposed to describe the round-trip range history accurately.The equivalent range model not only considers the curved trajectory,but also reduces the complexity of real range model.Compared with the other range models considering the errors of "stop-go",the equivalent range model has high precision and wide applicability.2.In order to solve the focusing problem caused by highly squint for spaceborne beam steering SAR(BS-SAR),a uniform Omega-K algorithm(?KA)is proposed.Firstly,the signal characteristics of different modes in the case of squinted is analyzed,including spotlight,sliding spotlight and TOPS,and a uniform signal model for BS-SAR is constructed.Then,the effects of highly squint angle and the variance of beam steering on the modulation rate and bandwidth of Doppler are studied.Based on the results,the traditional two-step processing is modified to make it more applicable to the highly squinted BS-SAR case.Finally,according to the property of Fourier transform,an accurate squinted two-dimension(2-D)spectrum expression after preprocessing is obtained,which is an accurate reference match function(RFM)expression for the squinted ?KA.Then,this RFM is used to focus processing.The proposed algorithm fully considers the effects of squint angle and beam steering variance,and can eliminate the spectrum aliasing and range-azimuth coupling.Moreover,it can process various mode of spaceborne and the procedures are simple and highly efficient.3.An azimuth frequency nonlinear chirp scaling algorithm(AFNCS)is proposed to process the case of high resolution LEO-SAR.With the increase of synthetic aperture time,the LEO-SAR moves along with curved path.This will result in two effects.On the one hand,the traditional range model is invalid;On the other hand,the performance of conventional algorithms,which are based on the assumption of rectilinear motion with constant velocity,is deteriorated.To solve these problems,a signal model of high resolution spotlight spaceborne SAR is made firstly.The equivalent acceleration vector is introduced to the range model to describe the actual range history accurately.Then,the 2-D spectrum based on the range model is derived by the method of series reversion(MSR),and the analyses of space-variant 2-D spectrum are performed.At last,the pre-filtering in azimuth time domain is integrated with azimuth frequency scaling to remove the space variance and range-azimuth coupling,and finish the focusing.This method can eliminate the azimuth space variance effectively and extend the scope of imaging.4.A range-azimuth 2-D scaling algorithm is proposed to deal with the range-azimuth 2-D space variance of GEO-SAR.Firstly,according to geometry of the Satellite–Earth and the physical knowledge,the actual trajectory of satellite can be described by the motion vectors,and the accurate instantaneous slant range expression can be obtained with the consideration of Earth rotation.Secondly,the quantitative analyses of range cell migration(RCM),the modulation rate of range and the azimuth modulation term of 2-D spectrum are performed.Eventually,based on the analyses of space-variance,the scaling method is modified to remove the range variance of RCM.The azimuth variance of azimuth modulation term is compensated by the proposed two-step scaling method,and the range variance is eliminated by updating along the range direction simultaneously.Then,the image processing can be finished.Because the effects of 2-D space-variance are fully considered and the most operations are made in range-Doppler domain,the proposed focusing method can process the large-scene echo data with high efficiency.