在干涉检测中,常用各种调制手段将被测信息加载到干涉图中,通过对检测得到的条纹图样进行分析,便可以得到被测信息。如果实验条件不允许施加任何调制手段,探测器检测得到的将很有可能是具有闭合条纹的单幅干涉图。正则化相位跟随(RPT)技术可以对单幅闭合干涉图进行相位重构,但其最主要的问题是精度过低。从最小二乘法角度出发,在经典RPT技术基础上,利用黄金分割(GS)算法对干涉图相位进行二次逼近,实现了对单幅闭合干涉图的高精度相位重构。计算机仿真结果表明该算法在保持了经典RPT算法各项优点的同时能够有效提高相位重构精度。对实际干涉图进行处理的结果与ZYGO GPI干涉仪进行了对比,结果较吻合,平均峰-谷(PV)值精度优于λ/20。

Different modulation methods can be adopted in interferometry and the desired information can be obtained with the interferograms. But sometimes the very nature of the experimental setup does not allow us to introduce any modulation method. Then we are forced to demodulate the interferograms with close fringes. Regularized phase-tracking (RPT) method can demodulate close fringes but with low precision. This paper presents a novel algorithm,called RPT approached & golden section algorithm (RPT&GS),to demodulate close fringes based on the RPT method. Starting from the result of classical RPT,we use the golden section algorithm to search for the accurate phase of single close fringe. Computer simulation shows the RPT&GS contains all the advantages of RPT besides high demodulation precision. Comparison experiments show good agreement with ZYGO GPI interferometer on real close fringes,and the average peak-valley (PV) value accuracy is better than λ/20.