中文摘要：

超高分辨率条件下，机载合成孔径雷达（SAR）发射信号带宽大，合成孔径时间比较长，对成像处理算法的精度和效率要求较高。现有近似频率域处理和时间域滤波反投影（FBP）算法聚焦SAR数据时均存在诸多问题。基于微局部分析方法，提出了一种新颖的频率域滤波反投影（FD-FBP）成像处理方案。首先，利用Keystone变换简化了数据距离多普勒（RD）域徙动表达式。然后，在RD域进行反投影操作，对参考位置处反投影数据进行移位、相位补偿和FFT等操作即可以得到图像，从而在保证算法精确性的前提下有效降低了运算效率，实现了频率域方法的高效率和时间域方法的精确性特点的结合。最后，点目标仿真和实测数据处理以及与FBP等算法的对比验证了该方法的有效性。

英文摘要：

In an ultra-high resolution airborne synthetic aperture radar （SAR）, both the bandwidth of the transmitted signal and the synthetic aperture time are increased, which results in higher requirements for accuracy and efficiency of the imaging algorithms. Many problems exist when using the approximated frequency domain algorithms and time domain filtered back- projection （FBP） algorithm to process SAR data. Based on microlocal analysis, this paper proposes a novel frequency do- main FBP （FD-FBP） algorithm. In this algorithm, Keystone transform is used to simplify the range cell migration expression of the SAR data in range Doppler（RD） domain first. Then, backprojection operation is implemented in RD domain, where shift, phase compensation and FFT operations are performed on the backprojected data of reference imaging points to obtain the imagery, which reduces the computational burden and realizes the combination of the efficiency of frequency domain al- gorithms and accuracy of the time domain algorithms. Finally, point target simulation and real-data processing result and their comparison with the FBP algorithm results validate this algorithm.