Research Of Image Formation Processing Techniques Based On Airborne UWB SAR Real Data

Airborne ultra-wideband synthetic aperture radar (UWB SAR) working in low frequency band can produce fine images of concealed targets with the capability of penetrating the foliage and the earth surface to reconnoitre wide-range scene, which is very valuable for military purpose. This dissertation work was supported in part by the National Defense Research Projects. According to the problems currently faced in the real data image formation processing on airborne P-band ultra wideband synthetic aperture radar (UWB SAR), this dissertation researched three subjects—wide-range scene imaging algorithm, motion compensation and image interferometry.On the first subject, the phase approximation of image formation algorithms for UWB SAR operating on P band was discussed in detail. The methods to determine the consistent focusing swath width in Chirp Scaling algorithm using residual coupling phases and range spectrum characteristics were also proposed. In order to make the theoretic results more realizable, the dissertation analyzed the formatting non-orthogonal sidelobes in UWB SAR images. The weighting function with variable length was employed to give a suppression of non-orthogonal sidelobes. By this method, echoes can be more tolerant to coupling phases in focusing so that the consistent focusing swath width was greatly increased. Using the above conclusions as guidance in segmented imaging, wide swath UWB SAR data had been successfully focused.In UWB SAR motion compensation (MOCO), the content was divided into two chapters, which concentrate on low-order MOCO and residual MOCO, respectively. To solve the range dependency of doppler parameters in UWB SAR subaperture model, an airborne motion error estimation algorithm was presented, in which doppler frequency modulation ratio was estimated by sub-range block data. Subsequently, these estimated values were jointed as a sparse matrix equation and the acceleration of direction Y was attained. Meanwhile, in real UWB SAR data processing, some questions from the doppler parameter estimation procedures was proposed and analyzed. And on the basis of the above work, the dissertation proposed several practical and important methods in map-drift algorithm realization and subaperture length calculation. To meet the feature of UWB SAR subaperture residual phase error model, an improved PWE-PGA (phase weight estimation - phase gradient autofocus) algorithm was proposed, whose intention was to get two range-independency factors. The effectiveness of the proposed methods had been verified with real UWB SAR data.In UWB SAR images interferometry, the dissertation first analyzed the advantages in UWB SAR interferometry (UWB InSAR) comparing to general narrow-band InSAR and concluded the difficulties in UWB InSAR realization. Afterward, the processing procedure of interferogram generation from two real UWB SAR images was demonstrated, which includes 2D-complex-images coregistration, pre-filtering and interferogram estimation. An interpolation algorithm based on local adaptive autocorrelation function estimation was proposed to appease complex-image interpolation. Based on the joint subspace projection estimating interferogram algorithm, a weighted joint autocorrelation matrix method was also proposed to improve the precision of interferogram estimation. In addition, the dissertation analyzed the time-varying baseline and deducted the offset formula of azimuthal spectrum of interferometric phase. Finally, the rationality of the interferometric phase fringes generated from those two real UWB SAR images was proved by the predicited results based on practical terrain.