基于非均匀快速傅里叶变换的SAR方位向运动补偿算法

沿航向的非匀速运动对SAR成像质量有很大影响，而运动补偿后SAR图像存在几何形变，影响子图像拼接和多波段SAR图像融合。当沿航向速度误差较大时，实图像域插值校正后图像残留的几何形变不能忽略。基于上述问题，该文提出一种基于非均匀傅里叶变换的SAR方位向运动补偿算法，算法直接对方位向的非均匀数据进行非均匀快速傅里叶变换(NUFFT)。该算法的定位误差和几何形变比实图像域插值校正几何形变算法小1到2个数量级，对沿航向速度误差有很强的鲁棒性，补偿后数据的幅度和相位信息都得以保留。SAR仿真和实测数据验证了该算法的有效性。

An Along-track Motion Compensation Algorithm Based on Non-Uniform Fast Fourier Transform (NUFFT) for SAR

The non-equispaced along-track motion has serious impact on SAR imaging. The SAR images are geometric distorted after the along-track motion compensation, which may impair the quality of the sub-image connection and the multi-spectral SAR image fusion. When the along-track velocity error is large the remained geometric distortion after the interpolation geometric distortion correction can not be ignored. In this paper, a SAR motion compensation algorithm is presented based on Non-Uniform Fast Fourier Transform (NUFFT) directly to the azimuth non-equispaced data. The algorithm has 1 to 2 quantitative reduction in location error and geometric distortion than interpolation geometric distortion correction algorithm and has high robustness on along-track velocity error. At the same time, the amplitude and phase information can both be saved based on the proposed algorithm. The results of the NUFFT compensation for simulated data and the real SAR data verify the effectiveness of the proposed algorithm.