Phase adjustment with the autofocus process is essential to inverse synthetic aperture radar (ISAR) imagery of manoeuvring targets. Conventional parametric and non-parametric autofocus methods usually degrade in front of error model mismatch and noise interference. Especially, the target manoeuvre also induces higher order phases that have the spatial-variant characteristic. In this study, a range-dependent semi-parametric autofocus algorithm for ISAR imagery is proposed. The range-variant phase is modelled semi-parametrically with a discrete cosine transform kernel. Implemented by the maximum sharpness optimisation, the algorithm well adapts to complex phase forms and enhances the robustness to strong noise. Accelerated solver to the optimisation is also introduced. Simulation and real data experiments confirm the advantages of the range-dependent semi-parametric autofocus algorithm compared with conventional algorithms.

中文翻译：

---

依赖于范围的半参数自动对焦，可操纵目标逆合成孔径图像

利用自动聚焦过程进行相位调整对于机动目标的合成孔径雷达（ISAR）逆成像至关重要。常规的参数和非参数自动聚焦方法通常会在误差模型失配和噪声干扰之前退化。尤其是，目标机动还诱发具有空间变化特征的高阶相位。在这项研究中，提出了一种针对距离的半参数ISAR图像自动聚焦算法。使用离散余弦变换内核对范围可变相位进行半参数建模。通过最大锐度优化实现，该算法很好地适应了复杂的相位形式，并增强了对强噪声的鲁棒性。还介绍了用于优化的加速求解器。