Rapid Determination of Moisture in GuoKui by Fourier Transform Near Infrared Spectroscopy

Fourier transform near infrared spectroscopy is used to establish an analysis model of moisture in GuoKui.Near infrared spectra of sixty-one samples is pretreated with First Derivative and Multiplicative Scatter Correction(MSC) to eliminate noise.The wavebands containing 7 501.9～4 597.6 cm-1 are selected and 10 numbers of factors are used.The calibration model between the Near-infrared spectra of GuoKui and the reference values measured by Chinese standard method is built with partial least squares.The correlation coefficients(R2) and the root-mean-square error of estimated(RMSECV) of the quantitative calibration model of moisture content in GuoKui are 99.03% and 0.409%,respectively.Besides,The external correlation coefficients(R2) and the root-mean-square error of prediction(RMSEP) attained by external validation of nineteen unknown samples,respectively,are 97.99% and 0.341%.The results indicate that the near infrared spectroscopy has high accuracy and can meet the rapid detection precision of moisture in GuoKui.