Robust Autofocusing Approach for Highly Squinted SAR Imagery Using the Extended Wavenumber Algorithm
Robust Autofocusing Approach for Highly Squinted SAR Imagery Using the Extended Wavenumber Algorithm
复制标题
DOI：
10.1109/tgrs.2013.2276112
复制
发表时间：
2013-10-01
影响因子：
8.2
通讯作者：
Bao, Zheng
中科院分区：
文献类型：
--
作者：
Xu, Gang;Xing, Mengdao;Bao, Zheng
对于大斜视合成孔径雷达(SAR)成像，在传感器轨迹为理想直线的情况下，波数域SAR处理算法被普遍认为是一种理想的SAR聚焦解决方案。然而，机载合成孔径雷达对大气湍流非常敏感，会导致严重的弹道偏差。本文提出了一种基于扩展波数算法的大斜视机载SAR图像的稳健自聚焦方法，该方法能够估计距离相关的相位误差。为了应用所提出的自聚焦方案，对圆锥参考系中的运动误差模型进行了详细的分析，给出了斜视SAR的距离相关相位误差公式。该方法分三步进行：1)针对大斜视SAR去斜视后不可避免的残余相位，提出了一种改进的斜视相位梯度自聚焦算法(SPGA)来恢复与距离无关的相位误差；2)基于建立的运动误差模型，利用局部最大似然加权SPGA算法估计剩余的距离相关相位误差；3)采用两步法进行运动补偿，分别得到与距离无关和与距离相关的校正。基于测量数据的实验表明，所提出的自聚焦方法对于大斜视SAR成像具有很好的性能。
For highly squinted synthetic aperture radar (SAR) imaging, the wavenumber domain SAR processing algorithm is commonly accepted as an ideal solution to SAR focusing in the case of an ideal straight sensor trajectory. However, airborne SAR is very sensitive to atmospheric turbulence that causes serious trajectory deviations. In this paper, we propose a robust autofocusing approach for highly squinted airborne SAR imagery using the extended wavenumber algorithm, being capable of estimating the range-dependent phase errors. To apply the proposed autofocusing scheme, a detailed analysis of the motion error model in the conical reference system is presented, where the formulation of range-dependent phase errors for squinted SAR is given. The proposed autofocusing approach is performed by a three-step process: 1) referring to the inevitable residual phase after deramping for highly squinted SAR, a modified squinted phase gradient autofocusing (SPGA) algorithm is put forward to retrieve the range-independent phase errors; 2) based on the established motion error model, the residual range-dependent phase errors are estimated using a local maximum likelihood-weighted SPGA algorithm; and 3) motion compensation is executed by a two-step approach to reach the range-independent and range-dependent corrections, respectively. Experiments based on measured data have shown that the proposed autofocusing approach performs well for highly squinted SAR imaging.