基于频域共轭梯度法的交替迭代复原算法研究

对迭代盲目去卷积复原方法进行了研究,提出了基于频域共轭梯度法的交替迭代优化复原算法,在频域上构造了关于目标图像和点扩展函数频谱的误差代价函数,将共轭梯度法引入到频谱误差代价函数的极小化过程中,并将空域非负性和频域带限等先验约束知识融合到对目标图像和点扩展函数的迭代交替优化估计过程中,取得了预期的复原效果,增强了算法的抗噪性和稳定性。在微机上进行了一系列的复原实验,实验结果表明算法复原效果好,抗噪能力强,速度较快,且能恢复具有复杂背景的目标图像。

Research on algorithm of alternant iterative blind restoration based on conjugate gradient method of frequency domain

The restoration algorithm of iterative blind deconvolution is studied in this paper,an ahernant iterative optimized restoration algorithm based on conjugate gradient on frequency domain is proposed.The error cost function of the frequency spectrum of the object image and the PSF in frequency domain is constructed,and the conjugate gradient method is introduced to minimize the error cost function,in which the constrains of the PSF values being non-negative and bandwidth limits are incorporated into the estimation process of the object image and the PSF, resulting in that good restorations are expected and that the abilities of robustness and noise-resisting of the proposed algorithm are enhanced.A series of restoration experiments have been performed on PC,showing that the proposed algorithm can obtained good restoration results and it has high level of noise resisting,and it can recover the object image with complex background.