Study On High Speed Maneuvering Platforms High Squint SAR Wide Swath Imaging Algorithm

Synthetic aperture radar(SAR)can get day-and-night and weather-independent images at a long distance.It has been applied to climate change research,earth monitoring and change detection.Arranging the radar on high speed maneuvering platform makes it possible for the platform to get radar images from both sides of the fight track.SAR imaging on high speed maneuvering platform has been one of the fundamental techniques for missile-borne radar seeker.This technique makes it possible for missiles to get real-time images of regions along the flight track which are useful for image matching guidance and targeting.Different from airborne platforms,the missile-borne platform has its own features,such as small size of antenna,high speed and curved trajectories.It also requires the ability of forward side-looking mode.These features result in several corresponding difficulties when dealing with missile-borne SAR imaging.The algorithm should be effective enough for real-time processing and able to process part aperture data.It needs to get an accurate range model for the platform motion and then handle the problem of strong range-azimuth coupling and azimuth ambiguous caused by beam steering.This thesis mainly aims to solve the above mentioned missile-borne SAR imaging problems,such as high accurate range modeling,envelope space-variance correction,phase space-variance correction and ambiguity resolution.(1)Frequency domain phase filtering for space-variant Doppler parameters correction Chapter 2 proposed a frequency domain phase filtering(FDPF)based imaging algorithm.In dealing with high squint mode SAR data,range walk correction(RWC)is common method to remove the range-azimuth coupling.It can also reduce the Doppler bandwidth of signal and the requirement of sampling rate.For the trajectory of the platform is no longer a straight line,the conventional range model based on hyperbolic equation is invalid.The higher order terms in the range formula make it hard to derive the 2-D spectrum of the signal.Some earlier methods have applied the method of series reversion(MSR)to derive the spectrum formula.But they neglected the space-variant of Doppler chirp rate which makes the focusing performance not that good.After them,some modified algorithms are made which try to remove the variant Doppler chirp rate terms based on the instantaneous Doppler relations.However,this method is only valid based on the far-field assumption which cannot be always satisfied in practical conditions.Some nonlinear chirp scaling(NCS)based algorithm can also be applied to solve this problem.But none of these methods can be directly applied to missile-borne SAR imaging.The proposed algorithm in this chapter is designed to process part aperture data and is effective for real-time on board data processing.The preprocessing is first made to compensate the acceleration phase terms.Then the azimuth signal can be uniformly focused after the FDPF for the space-variant term correction.(2)Orthogonal expansion range model and Doppler resampling methodChapter 3 proposed a modified imaging range model integrated with the Doppler resampling technique.Based on the analysis of signal characteristics,the signal is modeled based on the orthogonal expansion range model.This model is matched for the coordinate rotation after range walk correction(RWC)and has less approximation errors than the reference one.Different from the NCS method,the Doppler resampling can remove the space-variant Doppler chirp rate term without inserting other phase terms.In the experiment part,the proposed algorithm is proved to be more effective than the reference one.(3)Frequency domain backprojection for azimuth focusingChapter 4 proposed a frequency domain backprojection(FDBP)algorithm.A range model based on local Cartesian coordinate is proposed to handle the mismatch between the signal after RWC and the conventional range model.By the sub-region range migration correction(RCMC),the space-variant RCMC terms and the acceleration phase terms are all well compensated.Finally,the azimuth focusing is done by coherent integral in range-Doppler domain which is the procedure of FDBP.Compared with the conventional range model,the proposed one has less approximation errors.The processing of different sub-region signals can share a same processing architecture.All the sub-region images can be directly connected without further interpolation.This method has better azimuth focusing performance over the reference one.(4)Space-variant envelop correction based on Keystone transform and nonlinear deramp for beam steering mode imagingChapter 5 aimed to solve the imaging problems: range modeling,the correction of spacevariant range migration and space-variant Doppler chirp rate,in the beam-steering mode.After a slightly modification of range model in Chapter 4,targets with same reference range are defined on a same iso-range line,which is called the high squint concentric circle range model.In order to correct the space-variant components of RCMC terms,linear deramp and range-independent phase is first multiplied with the signal.After Keystone transform and phase compensation,the range migration can be uniformly compensated in time-domain.The Keystone transform of Doppler folded signal can be derived by extra two phase terms multiplication.Targets with same reference range are focused at the same reference range which is perfectly matched with the proposed range model.According to the time-frequency diagram(TFD)analysis,both the nonlinear variance and range-dependence of signal are taken into consideration during the azimuth signal reconstruction.The Doppler chirp rate variance is corrected in azimuth time domain by NCS method.This azimuth processing is based on the reconstructed range model.Finally,the signal is focused in 2-D time domain without zero-padding and sub-aperture processing.The simulated SAR data has verified all the key steps and the focusing performance.