Research on Calibration Transfer Method of Portable Near Infrared Instrument

摘要

The establishment of the near-infrared spectrum prediction model is usually affected by environmental factors such as humidity and temperature, and other uncontrollable factors, which makes the original model unable to predict new sample components, resulting in a decline in model prediction accuracy. This paper discusses the calibration transfer method of TI-NANO portable near infrared instruments. The slope/intercept correction algorithm and the direct standardization correction algorithm are used for research. The results show that for the slope/intercept algorithm, the RMSEP of the model is reduced by 0.825, 0.509, 0.706, and 1.378 respectively, and R2 is increased by 0.133, 0.326, 0.116, and 0.288, respectively; for the direct standardization algorithm, the RMSEP of the model decreased by 0.547, 0.303, 0.897 and 1.633 respectively, and R2 increased by 0.195, 0.322, 0.135 and 0.204 respectively. After being processed by the calibration transfer method, the accuracy of the prediction results has been improved, especially the direct standardization method is preferable.