摘要
首先介绍了相位恢复技术中的Gerchberg-Saxton算法和梯度搜索算法。推导了当任意多帧输入图像及它们的离焦量作为输入时,梯度搜索算法的目标函数分别关于广义光瞳、波前以及泽尼克系数的偏导数。揭示了GS算法与梯度搜索算法之间的关系。针对单幅和多幅图像作为输入时分别用GS算法和梯度搜索算法设计了仿真实验,实验结果显示对于单幅图像作为输入时,梯度搜索算法明显优于GS算法。对于多帧不同离焦量的图像作为输入时,GS算法和梯度搜索算法都能很好的解算出波前,但梯度搜索算法的收敛速度明显优于GS算法。
This paper introduces two algorithms of the phase retrieval——Gerchberg-Saxton algorithm and gradient search algorithm.We respectively get the objective function of gradient search algorithm about the generalized pupil,wave-front and the Zernike coefficients of the partial derivatives when multi-frame images and their defocus are used as the inputs.The relationship between GS algorithm and the gradient search algorithm is revealed.Simulation experiments are carried out with GS algorithm and gradient search algorithm when single and multiple images are used as inputs.The experiment results show that the gradient search algorithm is superior to GS algorithm for single image input.Both GS algorithm and gradient search algorithm can work out wave-front for multi-frame images with different defocus as the input,but the convergence speed of gradient search algorithm is evidently better than GS algorithm.
出处
《激光与红外》
CAS
CSCD
北大核心
2012年第2期217-221,共5页
Laser & Infrared
基金
国家"863"高技术研究发展计划项目(No.2009AA8080603)资助
关键词
相位恢复
GS算法
梯度搜索算法
物域
傅里叶域
phase retrieval
GS algorithm
gradient search algorithm
object-domain
Fourier-domain
