为了提高自适应光学图像复原效果, 提出了一种新的多重约束非对称图像迭代盲解卷积算法。首先, 在点扩散函数(PSF)频率域引入带宽有限约束来提高迭代盲解卷积算法的可靠性; 然后, 在PSF空间域引入支持域动态更新的思想以加快迭代盲解卷积算法收敛速度; 最后, 自动计算迭代盲解卷积算法的非对称因子以提高算法的自适应性。模拟实验结果表明, 与RL-IBD算法比较, 新算法迭代次数减少22.4%、峰值信噪比提高10.18 dB。在FK5-857和某双星的自适应光学图像复原实验中, 也取得很好的复原效果。

A novel multi-limit unsymmetrical iterative blind deconvolution(MLIBD) algorithm was presented to enhance the performance of adaptive optics image restoration. The algorithm enhances the reliability of iterative blind deconvolution by introducing the bandwidth limit into the frequency domain of point spread function(PSF), and adopts the PSF dynamic support region estimation to improve the convergence speed. The unsymmetrical factor is automatically computed to advance its adaptivity. Image deconvolution comparing experiments between Richardson-Lucy IBD and MLIBD were done, and the result indicates that the iteration number is reduced by 22.4% and the peak signal-to-noise ratio is improved by 10.18 dB with MLIBD method. The performance of MLIBD algorithm is outstanding in the images restoration the FK5-857 adaptive optics and the double-star adaptive optics.