低信噪比、低对比度SAR图像自聚焦处理方法研究

In order to obtain high resolution and high quality synthetic aperture radar (SAR) images, the autofocus algorithms must be used to estimate and correct the phase error. Therefore, the technology of autofocus algorithms has been one of the hot points in the SAR singal processing research field. With the further development of technology, people have put forward higher requirement for the autofocus algorithms, especially at low signal-to-noise ratio (SNR) and low-contrast. This program will present a novel autofocusing algorithm using the signal subspace by the simultaneous processing of multiple-pulse vectors of range-compressed data, which will have better performance than the traditional autofocusing algorithm. In order to meet the requirement of algorithm complexity, robustness and accuracy, some key technologies have to be solved. This program will make a thorough research about the signal subspace estimating method, subaperture processing method and the subaperture phase error combining technology. First, the approximation algorithms of signal subspace will be used to reduce the computational cost of the autofocusing algorithm. Then, the multi-subaperture processing technology will be presented to solve the problem of limited training samples in autofocusing algorithm to enhance robustness. However, subaperture autofocusing also incurs the problem of coherent subaperture phase error combining. To solve this problem, the map drift (MD) technique will be employed to decreasing the accumulation of estimation errors during the procedure of the subaperture phase error combining. This research production can be applied on the autofocusing of the SAR images with low-SNR and low-contrast, and can be used for the development of the theory of SAR imaging and motion compensation.

为了获得高分辨率、高质量的SAR图像，须利用自聚焦处理估计相位误差并进行补偿。因此，自聚焦处理一直是雷达信号处理研究的热点之一。现有自聚焦方法对图像信噪比要求较高。然而，随着SAR硬件和信号处理技术的发展，要求自聚焦方法可以在低信噪比、低对比度条件下，依然保证对于相位误差函数的高精度估计。在本项目中，申请者将在已有工作基础上，结合信号子空间处理技术，基于多脉冲联合估计的方法，实现对低信噪比、低对比度SAR图像自聚焦处理。针对该信号处理方案的复杂度、鲁棒性和估计精度三方面，开展信号子空间估计方法、子孔径处理技术和子孔径相位误差函数拼接技术研究，研究目标是探索低信噪比、低对比度SAR自聚焦处理的高精度和高效率处理流程。本项目研究成果是对经典SAR自聚焦处理的改进，使其从计算速度和精度两方面同时得到提高，可应用于各种恶劣条件下所获得的SAR图像中，是对SAR成像成像和运动补偿理论的发展和完善。

本项目为实现低信噪比、低对比度SAR自聚焦处理，围绕该信号处理方案的复杂度、鲁棒性和估计精度三方面，开展信号子空间估计方法、子孔径处理技术和子孔径相位误差函数拼接技术研究，主要完成的研究内容包括以下六个方面。（1）从信号子空间角度出发，提出利用PAST算法避免特征分解过程，并从自聚焦处理角度优化算法流程，有效降低算法复杂度，提高自聚焦处理实时性。（2）引入子孔径处理技术，研究在有限距离单元数条件下，参与相位估计脉冲数对自聚焦算法性能的影响，提出合理的子孔径脉冲数和子孔径间重复脉冲设置方案，提高自聚焦处理鲁棒性。（3）研究高精度相位误差函数拼接问题，提出利用MD算法实现高精度相位误差函数拼接，并针对不同SAR成像模式提出了不同的相位误差函数拼接处理方案，保证自聚焦处理精度。（4）基于自聚焦处理原理，延伸研究多基线InSAR干涉相位图滤波技术，提出了基于信号子空间估计的多基线InSAR干涉相位图滤波方法，有效解决了较长基线对应干涉相位图滤波问题。（5）基于SAR系统和WSN系统的相似性，延伸研究WSN系统节点定位技术，提出了基于信号相位搜索的WSN节点定位算法，有效提高了WSN节点定位精度。（6）基于相控阵雷达和大规模MIMO系统的相似性，延伸研究大规模MIMO系统波束成形方法，提出了一种毫米波高铁车地通信的自适应多波束成形与波束切换方法，显著提升高铁无线通信系统的容量，同时降低整个系统的通信中断概率。项目组完成了预定的研究任务，相关研究成果是对SAR成像和运动补偿理论的发展和完善，并可以延伸服务于多个相关领域。