为了标定液晶相位可变延迟器(Liquid Crystal Variable Retarder, LCVR)的相位延迟特性, 在25 ℃、405 nm波长下, 利用搭建的测量装置采集了141组实验样本, 其中71组样本为训练集, 70组样本为预测集, 利用最小二乘支持向量机(Least Squares Support Vector Machines, LSSVM)和支持向量机(Support Vector Machines, SVM)算法建立LCVR相位延迟量和驱动电压相关数学模型。实验表明, GASVR、PSOSVR、LSSVM方法下最大波长偏差?驻λ分别为0.013 6λ、0.013 7λ和0.004 5λ, 均方误差提高两倍, 通过比较, 说明该模型能快速准确地预测LCVR工作范围内全部波长、全部电压值下的相位延迟。该方法可作为LCVR相位延迟特性标定的有效手段。

To calibrate the phase characteristic of a Liquid Crystal Variable Retarder(LCVR) attached with a compensator. For LCVR, under 25 ℃, 405 nm wavelength, the total of 141 experimental samples were obtained by the built setup based on Stokes parameters, in which 71 samples(calibration set) were used for model calibration, and 70 samples(prediction set) were used to validate the calibrated model independently. Least Squares Support Vector Machine(LSSVM) and Support Vector Machines(SVM) was used to calibrate model. Test results show, for prediction model based on GASVR, PSOSVR, LSSVM, the maximum wavelength deviation are respectively 0.013 6λ、0.013 7λ and 0.004 5λ, the mean square error is increased by 2 times, and the phase retardation of all the wavelengths and all the voltage values of LCVR in the effective working range could be rapidly and accurately obtained by the LSSVM model, thies the prediction model based on LSSVM was an effective mean for LCVR phase calibration.