基于TIE和加速角谱迭代算法的二维相位恢复  

Two-Dimensional Phase Retrieval Based on TIE and Accelerate Angular Spectrum Iteration Algorithm

作　　者：梁丽[1,2] 杨玲[1,2] 王中科[3] 甄小琼[1,2] 

机构地区：[1]成都信息工程学院电子工程学院,四川成都610225 [2]中国气象局大气探测重点开放实验室,四川成都610225 [3]成都信息工程学院网络工程学院,四川成都610225

出　　处：《激光与光电子学进展》2013年第2期139-143,共5页Laser & Optoelectronics Progress

基　　金：公益性行业(气象)科研专项(GYHY201106047)资助课题

摘　　要：在Gerchberg-Saxton(GS)算法的基础上,引入光强传播方程法(TIE)和加速角谱迭代算法,提出了基于TIE的加速角谱迭代算法,实现了更加精准快速的相位恢复。该算法通过3个面的光强(一个输入面,两个输出面)进行相位恢复,首先通过光强传播方程得到初始的相位,进而使用加速角谱迭代算法进行进一步的迭代恢复相位。通过数值仿真得出,该算法在二维图像相位恢复过程中更加精准快速,抗噪性更强。在可恢复相位的控制范围内,迭代40次即可趋于稳定,均方根(RMS)误差可控制在10-6数量级。Based on conventional Gerchberg-Saxton （GS） algorithm, transport of intensity equation （TIE） method and accelerated angular spectrum iterative algorithm are introduced to put forward a kind of more accurate and rapid phase recovery algorithm： accelerated angular spectrum iterative algorithm for two-dimensional （2D） phase retrieval from TIE. The intensity distribution of three planes, e.g., an incident plane and two emergent planes, is employed to retrieve the initial phase by TIE and then retrieve the phase of incident optical field by accelerated angular spectrum iterative algorithm. The simulation result shows that this algorithm can retrieve the phase distribution of an arbitrary 2D incident optical field in high accuracy. Through numerical simulation, this algorithm is proved in 2D image phase recovery process to be more accurate and faster and its noise immunity is stronger. In the recovery phase control range, the iteration tends to be stable after 40 steps, and the mean square （RMS） error can be controlled in the order of magnitude of 10-6.

关 键 词：图像处理 光计算 相位恢复 光强传播方程 加速角谱传递函数 

分 类 号：O436[机械工程—光学工程]