Calibration optimization for rice apparent amylose content by near infrared reflectance spectroscopy (NIRS)

Abstract

Using ground milled rice as test example, various methods for spectra pretreatment and different regression method were compared to form a good regression equation for rice apparent amylose content (AAC) by near infrared reflectance spectroscopy (NIRS).The results were as follows: Scatter correction for spectra had insignificant effects on calibration, and Standard normal variate (SNV) could be used in option; The first derivative is better for mathematical treatment of the spectra value than second or third derivative, and the suitable Segment Gap combination was “1, 5, 5,1” for data collection; Regression method had obvious effect on calibration, and the best one was Modified Partial Least Square (MPLS) method. In summary, the optimal calibration procedures could be “SNV/1, 5, 5, 1/MPLS” for NIRS analysis of rice AAC when using ground milled rice as test sample. When using such calibration parameters, an ideal regression equation was obtained with a SEP (Standard Error for Performance) as low as 0.84 and a R 2 (Determination coefficient for validation) as high as 94%.