Airborne Synthetic Aperture Radar Imaging And Research On Motion Compensation

Synthetic aperture radar(SAR)is an active imaging radar with all-day and allweather capabilities.SAR system outputs the high-resolution images which are the very foundation of the application of topographic measurement,target detection and explanation and environment survey as well.But,the echoes of airborne-SAR are easily contaminated during acquisition by the motion error which causes the rangecompressed signal an extra position shifts in range direction.In azimuth direction motion errors widen targets' mainlobes or induce paired echoes or submerge the targets into clutter.Consequently,the basic imaging procedure should take into account the case in the presence of motion error or it should combine the classic image formation algorithms with the motion compensation algorithms,which could bring solid foundation for SAR's popularization.The main problem in the airborne SAR motion compensation domain is how to estimate the motion error with high accuracy and solve its range-dependence(i.e.,range-variant characteristic),azimuth aperture-dependence(i.e.,aperture-variant characteristic)and time-dependence.As for the first one,most algorithms are based on azimuth phase gradient autofocus(PGA)and partly parametric autofocus;For the features of the motion errors,some of current algorithms focus on the azimuth timedependence(Two-step motion compensation),some of them deal with the azimuth aperture-dependence(time-division compensation,frequency-division compensation etc.),some solve the range-dependence(one-step motion compensation).This paper studies the estimation and compensation of motion error with precisely block processing.First,on the presence of residual motion error,this paper presents an image formation algorithm based on the automatic range cell migration correction and an adaptive pulse compression based on the fractional fourier transform.One,this paper considers the each echo's migration as variables,together with the weighted entropy of the corrected signal as the cost function and derives the precise close form of the gradient.Compared to the traditional residual-motion-error-parameterized methods,the automatic range cell migration correction has better robustness.Two,for the influence of the residual motion error on the azimuth phase history,this paper views the rotation angle of the fractional fourier transform as a variable and applies the stochastic gradient-descent to obtain the optimal angle with the entropy of compressed signal as the cost function,which can avoid the miss-match phenomenon effectively.Second,this paper proposes an autofocus algorithm based on the modified entropy and residual entropy(MERE)and optimal transfer method(OTM).First,under the analysis of the performance of the image's entropy,the paper defines the modified entropy and residual entropy of a SAR image and makes them weighted summation to construct the new cost function MERE that can not only strengthen the prominent targets but also prevent the energy of weak targets from leaking to the strong ones,which could ensure the correct structure of the scene(image).Second,this paper applies the optimal transfer method and designs the surrogate function of MERE in order to lower the calculating complexity.In addition,an adaptive order selection is also included to ensure that no over fitting or under fitting occur.The real airborne SAR data validates the effectiveness of the proposed algorithms.Third,a precise azimuth autofocus algorithm is carried out based on correlate input PGA and weighted total least square(WTLS).For the low signal-clutter-ratio(SCR)of traditional autofocus under complicated flying path,the paper collects the azimuth phase history before range cell migration correction,then applies correlate processing to the signal with different range cell and next make these correlated signal as the input of PGA so as to improve the input's SCR and precision of PGA.Second,The whole autofocus process is consisted of coarse estimate stage,fine estimate stage and linear term estimate stage.The WTLS is applied in each stage to obtain the corresponding phase error.These strategies could ensure the estimate precision no matter under complicated motion errors or slowly varying motion errors.In the stage of experiments,we use the simulated and real data to proof the validation of the proposed algorithms.Forth,this paper proposes motion compensation algorithms with capacity of tackling the range-dependence and aperture-dependence.Based on the analysis of the geometry relation among motion error,range and instantaneous slant angle,the paper can compensate the motion error completely.First,expand the motion error in range direction with Taylor series and apply the chirp scaling to compensate the range-dependent component.Next,derive the corresponding two precise relations between Doppler frequency and instantaneous slant angle under two-step motion compensation and chirp scaling processing,respectively to compensate the residual aperture-dependent motion error.For the case of dechrip receiving,a frequency scaling is investigated to spread the range-dependent envelope correction to the frequency-modulated continuous wave system.