光学多孔径共相综合是获得高分辨率成像的有效方法。针对子孔径间的活塞误差，提出一种基于干涉图像清晰化的校正方法。在理论分析的基础上，建立数值仿真模型，分析所提方法的校正范围和误差。结果表明，在复色光情况下，对于小于相干长度的活塞误差，以远场干涉图像的清晰度为目标函数、0.5个中心波长的整数倍为扰动步长，采用双边扰动梯度下降算法进行迭代优化，能够实现活塞误差的高精度闭环校正。

Co-phasing alignment of the optical multi-aperture is effective method in achieving high-resolution imaging. A correction method based on interference image sharpness is proposed to resolve the piston error between sub-apertures. A numerical simulation model is established and the correction range and piston error of the proposed correction method are analyzed according theoretically. According to the results, under polychromatic light, this iterative optimization method achieves a high-precision, closed-loop correction of piston errors smaller than the coherence length. In this method, the sharpness function of the far-field interference image is considered to be the objective function, and the integer multiple of the half-center wavelength is used as the step size for the bilateral perturbation gradient descent algorithm.