Research On Airborne SAR Three-dimensional Imaging Theory And Key Technologies

The three-dimensional (3-D) Synthetic Aperture Radar (SAR) technique is asignificant advancement of two-dimensional SAR. The3-D SAR approach solved theambiguity problems in the height direction of two-dimensional imaging byreconstructing the spatial distribution of observation area targets. It is attractingincreasing attention, due to its broad applications in key regions reconnaissance,topographic surveying, mapping and environment protection. In recent years,substantial3-D SAR research has occurred with advancement in imaging theory,imaging algorithm and experiment system. Unfortunately, the3-D SAR techniquecontinued to have problems hindering its extended application in radar imaging field,such as improving the resolution of3-D imaging system, improving the efficiency ofimaging processing in consideration of the imaging precision and designing experimentor practical radar system to meet the demands.This dissertation advances the research on SAR tomography and circular SAR withfocuses on3-D SAR imaging system theories, imaging methodologies and the keytechnologies based on the airborne platform. The main work and contributions are asfollows:1. The mechanism of multi-baseline airborne SAR tomography3-D imagingcapability is analyzed, and the imaging system model is built. Using mathematicalmethod, the basic index and conditions of imaging within this model are determined.The dissertation introduces multi-baseline SAR tomography3-D imaging methodsbased on the spectrum analysis and spectrum estimation. Through theory analysis andsimulation, the effects of sparse baseline conditions are studied.2. On the basis of analyzing imaging methodologies based on Singular ValueDecomposition (SVD) and Compressive Sensing, novel effective approachs forenhancing SAR image quality are proposed under the condition of sparse baselines. Thedissertation models research on the Sparse Bayesian Learning imaging algorithmcombined with Trench-Zohar algorithm and discusses their significance. Simulationresults and the comparison of the SVD algorithm demonstrate that this approach has better capability than the SVD algorithm with super resolution characteristics. It alsocarries out research on Orthogonal Matching Pursuit and Regularized OrthogonalMatching Pursuit algorithms in the application of sparse baseline SAR tomography. Thesimulation experiments demonstrate the feasibility of the algorithms, and theiradvantages on computation efficiency are then analyzed.3. On the basis of analyzing the echo signal model of the circular SAR, the3-Dimaging resolution with full aspect angel observation is modeled by the analysis of thesystem ambiguity function, and the relationship between imaging coherent angel andresolution is then studied. These models establish the foundation for the research onairborne circular SAR imaging processing methodology.4. In this dissertation, a detailed analysis of two kinds of airborne circular SAR3-D imaging processing methods is proposed. They are the traditional back-projectionalgorithm and a kind of3-D imaging algorithm based on wavefront reconstruction witheliminating trajectory correlative phase in angle frequency domain. With the goal ofsolving the bad imaging quality problem caused by the effect of trajectory error inpractical circular SAR system, the dissertation launches an error compensation approachcombined with subregion segmentation and error extraction of the scene center. Usingthe simulation data and two sets of experimental data, imaging experiments areperformed. The results demonstrate the advantage of imaging precision for the imagingalgorithm based on back-projection, and the great advantage of operation efficiency forthe imaging algorithm based on fast wavefront reconstruction. The effectiveness of theproposed error compensation method for trajectory error is consequently proved.Based on the research of this dissertation, the following innovative results areproduced:1. Based on Trench-Zohar algorithm and Sparse Bayesian Learning algorithm, ahighly efficient imaging method of SAR tomography with no compromise on imagingaccuracy is proposed. The imaging simulation experiments and comparative analysisdemonstrate the advantage of the algorithm on the imaging performance in the sparsebaseline conditions.2. A sparse baseline SAR3-D imaging algorithm based on Regularized OrthogonalMatching Pursuit is proposed. The simulation results demonstrate the advantage of thealgorithm on the amount of computation required and the reconstruction accuracy. 3. A fast wavefront reconstruction algorithm based on eliminating trajectorycorrelative phase is proposed. Angular frequency domain processing is utilized toeliminate trajectory correlative phase items. By doing so, the complex Hankel functioncalculations required in traditional wavefront reconstruction algorithm are avoided. Theeffectiveness of the algorithm and its advantage over the back-projection algorithm inthe amount of calculation are proved with simulations and actual data imagingexperiments.4. A trajectory error compensation approach combined with subregionsegmentation and error extraction of the scene center is proposed. Based on the fastwavefront reconstruction algorithm, this compensation approach utilizes error extractionof the scene center and subregion segmentation to narrow range of errors, and realizescommon airborne platform trajectory error compensation. It is a worthy attempt in usingthe circular SAR imaging processing in engineering.