基于多面循环Gerchberg-Saxton算法相位复原研究

为了避免传统Gerchberg-Saxton算法易陷入局部极值、而不能达到理想精度的问题，采用一种简单的循环方法，通过增加信息量，即增加一个平面，使算法能够比较有效地跳出局部极值，并进行了模拟计算。在光瞳平面振幅未知的情况下，利用焦平面或附近位置处的光强信息，对光瞳平面相位和振幅分布进行了恢复。结果表明，新算法达到较高的相位复原精度，进而实现了任意位置处光场的准确复原。

Research of phase retrieval based on multi-face circulatory Gerchberg-Saxton algorithm

To avoid the problem that traditional Gerchberg-Saxton (GS) algorithm was easy to get into a partial extremum and could not achieve the desired accuracy, a simple circulatory GS algorithm was used by increasing the amount of information, in another words, increasing one plane so that the GS algorithm could get out of the partial extremum effectively. After the simulative calculation, the phase and amplitude distribution of pupil plane was recovered by the intensity information of the focal plane or nearby position although the amplitude of pupil plane was unknown. The results show that the new algorithm can reach higher accuracy of phase retrieval and lead to precise retrieval of optical field on any planes.