Research On The Technology Of The Autofocus Method And Its Application In The Airborne Synthetic Aperture Radar Interferometry

Based on the relative motions between the antenna and target, synthetic aperture radar (SAR)could form a larger equivalent aperture antenna to obtain high azimuth resolution imageries bysignal processing techniques. It can significant enhance the ability of the long-distanceobservation for its all weather and all time character. After the developments in the past decades,the SAR imaging technologies become mature gradually. Because SAR is a coherent imagingsystem, there is an intimate connection between the image and the phase of the received signal. Inorder to obtain high resolution and high quality SAR images, the autofocus algorithms are used toestimate and correct the phase error. The technology of the interferometric synthetic aperture radar(InSAR) was used to obtain the information of the three-dimensional terrain since late1960’s. Andthe high quality SAR image is essential to the signal processing of InSAR elevation measure.Chapter1is the introduction of the dissertation. The history of SAR and InSAR technology isoutlined, respectively. And the applications and the developments of autofocus algorithms andInSAR technology are introduced in detail. At the end of this chapter, the aim and contents of ourwork are addressed.Chapter2studies a multi-subaperture autofocus algorithm. Due to the characteristics of theconventional strip-mapping mode SAR imagery, the present techniques of autofocus cannot bedirectly used to estimate the phase error functions. By using the autofocus algorithm forsubimages and the subaperture phase error combining with technology that based on the seconddifferences, we can achieve the phase error functions of the strip-map SAR imagery throughappropriately changing the construction of the original range-compressed data. However, thismethod will lead to the accumulation of estimation errors during the procedure of the subaperturephase error combining. To address these shortcomings, we present a multi-subaperture autofocusalgorithm, which combines the techniques of phase gradient autofocus (PGA) and map drift (MD).Comparison on the accuracy of this approach against the traditional algorithm is presented.Experimental results indicate that the presented PGA-MD methodology can improve the quality ofstrip-map SAR image effectively. In the spotlight mode, the subaperture-images represent exactlythe same scenario, and are highly similar after refocused by PGA. Therefore any of thesubaperture-image pairs can be used as input to MD, and the redundancy of multiplecross-correlating results serves to suppress the effects of noise and target scintillation. What’smore, it smoothly incorporates the estimation of residual range cell migration (RCM) to furtherimprove the quality of SAR image.Chapter3proposes a novel autofocus algorithm using the projection approximation subspacetracking (PAST) approach. SAR is a coherent imaging system. The key of obtaining highresolution and high quality SAR image is the maintenance of accurate coherent phases via usingautofocus algorithms. An eigenvector method for maximum-likelihood estimation (MLE) of phaseerrors for use in an autofocus algorithm for SAR imagery by the simultaneous processing ofmultiple-pulse vectors of range-compressed data has better performance than the algorithm ofPGA. However, this method requires eigendecomposition of the sample covariance matrix，whichis a task that is computationally expensive and limits the real-time application. In order toovercome this difficulty, a novel autofocus algorithm using PAST is presented. With thismethodology, the computational cost can be reduced effectively to the level of PGA via avoidingthe procedures of covariance matrix estimation and eigendecomposition. Monte Carlo tests and real SAR data processing validate that the new approach outperforms the mostly used PGA. Theperformance of multi-subaperture autofocus algorithm, which is detail described in the previouschapter, can be further improved by utilizing the PAST based autofocus algorithm instead of PGA.Chapter4introduces the crucial steps of InSAR technology. Analysis and research on thestatistical characteristic of interferometric phase is given by formula deducing. And then the signalprocessing of InSAR elevation measure is briefly presented. We employ the real SAR data toobtain the processing results of the partial key points of InSAR technology. This chapter alsodetailed analysis the reason of the target displacement in digital elevation model (DEM), andshows the mathematical relation between the target displacement and the SAR imaging geometryand the height of target. The simulation results validate the analysis.Chapter5focuses on the phase unwrapping. We investigate the Goldstein branch-cut phaseunwrapping method and the least squares estimation technique. And then a two-dimensional phaseunwrapping approach using equivalent residues is proposed for InSAR. In this proposed algorithm,the relationship between quality map and residues is used to find out the low quality unreliableregions, which are residues dense distribution and regarded as equivalent residues. Then, differentphase unwrapping strategies are taken for different quality regions. With this methodology,integration path crossing of unreliable regions, which may produce a phase error that propagatesto all the pixels in integration path, is prevented because that the unreliable regions are treated asequivalent residues. Each pixel inside equivalent residues is unwrapped based on its unwrappedneighbors, which breaks the limit of the absolute value of phase gradient between two adjacentpixels. Simulated and real SAR data processing validate the new approach.Chapter6summarizes the major contents in this thesis and points out the direction and thefocus of next research work in the future.