Influence of temperature on calibration transfer of near-infrared spectroscopy detection model for fructose content
The method of Slope/Bias correction was applied to study the calibration transfer of the fructose content detection model at different sample temperatures between two different fourier-transformation near-infrared spectrometer machines. The effect of the sample temperature on the result of calibration transfer was discussed. When modeled directly on the master and slave instruments, the root mean square of prediction error (RMSEP) of the prediction set each had a rising trend with the increase of temperature. But the change was not significant. When predicted directly, the results all were greater than 0. 86, thus being unsatisfactory. After transfer by S/B algorithm, the RMSEP was improved. The transfer result between the same temperature spectra was better than between the different. When transferred between the same temperature spectra, the results of 15℃,22 ℃ and 28 ℃ were 0. 317, 0. 389 and 0. 416, the difference being significant. It is shown that temperature variation has great influence on model transfer. With consistent sample temperature on the master and slave instruments ensured, it is good to choose an appropriate temperature for achieving the best transfer result.
near-infrared spectroscopycalibration transferfructose contentcalibration transfer
国家科技期刊平台
NSTL
万方数据
维普
