合成孔径雷达四维稀疏成像模型及处理方法研究

Synthetic aperture radar 4-D sparse imaging is an advanced subject in the field of the microwave remote sensing, which is a natural extension of the traditional microwave imaging technique. It allows measuring the scattering distribution in the azimuth-range-height-deformation velocity 4-D space and is fruitfully applied for detecting and monitoring the natural calamity, the subsidence of the ground and so on. The observation data obtained from 4-D synthetic aperture radar imaging system is sparse and non-uniform in the baseline-time plane. Hence, the imaging results are limited by high sidelobes. To overcome the resolution and sidelobes problem suffered from the traditional imaging methods, the theory of sparse signal processing is introduced for the 4-D synthetic aperture radar imaging in this project. The main contents of the project include three aspects. Firstly, the sparse representation methods for the different imaging scenes and the sparse scheme for the platform distributions are analyzed; the cost function based on sparsity constraint for the 4-D synthetic aperture radar imaging system is built, and the 4-D sparse imaging model is constructed. Secondly, the robust 4-D imaging methods based on the theory of sparse signal processing and optimization are investigated. Finally, the performance of the proposed algorithms for 4-D synthetic aperture radar sparse imaging system are studied; the influence of the parameters to the imaging processing is analyzed, and the effectiveness of the proposed methods is confirmed. Through the studies in the project, a methodical framework for 4-D synthetic aperture radar sparse imaging processing is expected to form, which should build the theoretical and technical foundations for the application of 4-D synthetic aperture radar sparse imaging technique in the civil and military areas. Consequently, this project has the important theoretical significance and application prospect.

合成孔径雷达四维稀疏成像技术是微波遥感的前沿研究课题，是传统微波成像技术的发展，具有方位-距离-高度-形变速率四维成像能力，在灾害和环境监测、地表沉降等领域具有重要的应用价值。本课题申请拟将稀疏信号处理理论和合成孔径雷达四维成像技术相结合，解决传统谱估计法和逆问题法在合成孔径雷达四维图像重建中存在旁瓣高、成像效果不佳的问题。课题的研究内容包括分析不同成像场景的信号稀疏表示策略和平台轨迹分布的稀疏化方案，构建四维合成孔径雷达系统的稀疏约束优化代价函数，建立合成孔径雷达四维稀疏成像模型；采用稀疏信号处理和最优化方法，研究稳健的合成孔径雷达四维稀疏成像处理方法；深入研究四维稀疏成像算法的性能，分析各参数对成像性能的影响，验证算法的有效性；形成一套完整的合成孔径雷达四维稀疏成像处理的理论框架，奠定合成孔径雷达四维稀疏成像技术在民用和军事领域应用的理论和技术基础。课题具有重要的理论意义和应用前景。

合成孔径雷达四维稀疏成像技术是微波遥感的前沿研究课题，是传统微波成像技术的发展，具有方位-距离-高度-形变速率四维成像能力，在灾害和环境监测、地表沉降等领域具有重要的应用价值。本课题对四维合成孔径雷达系统的成像几何和信号特征进行了深入分析，引入矢量化算子Vec和Kronecker乘积操作将合成孔径雷达四维成像处理转化为从噪声测量中对稀疏信号进行重构，构建了四维合成孔径雷达稀疏约束优化代价函数；对信号的稀疏先验分布进行了研究，提出了一种基于弱lp模型的信号可压缩性分析方法，为压缩重建时选择合适的信号先验分布模型提供了依据；针对传统正交基函数与信号特征不完全匹配时存在误差的问题，利用雷达平台的稀疏结构与信号特征构建了冗余字典，更加准确的描述信号的本质特征；针对单纯随机采样存在轨迹分布不均，成像效果不理想的问题，将具有蓝色噪声频谱特性的泊松碟采样方法扩展到多维空间，对四维合成孔径雷达的基线进行优化设计，结果表明分布更均匀的随机采样方法能够获得更好的测量矩阵性能；分析了噪声对重建方法的影响，结合贝叶斯估计和压缩感知理论，提出了基于贝叶斯压缩感知的合成孔径雷达四维成像方法，建立了稀疏有效的贝叶斯估计模型；为了提高重建精度，在成像过程中考虑相位的影响，将合成孔径雷达四维成像问题转化为目标复散射系数的幅度和相位联合重建问题，并通过在成像过程中引入相位信息来改善成像质量。同时对目标的点散射、区域散射特征等先验信息加以约束；分析了四维合成孔径雷达轨迹的随机分布、轨迹数量和噪声等参数对成像性能的影响，并通过实验验证了提出方法具有更好的抗噪性和稳健性。形成一套完整的合成孔径雷达四维稀疏成像处理的理论框架，为合成孔径雷达四维稀疏成像技术在民用和军事领域应用奠定了理论和技术基础。课题具有重要的理论意义和应用前景。