Quantitative analysis of melamine based on terahertz spectral data fusion
In order to meet the demand for accurate quantification of the illegal additive melamine in milk powder,terahertz time-domain spectroscopy was employed to measure the absorbance spectra of milk powder doped with melamine. The absorption spectra of mixtures of milk powder and melamine (concentration gradient from 0％ to 20％) and the pure substances were recorded in the range of 0.5 to 2.5 THz. The Savitzky-Golay first-order smoothing method was utilized to eliminate noise from the absorption spectra and to derive the corresponding derivative spectra. Combining chemometric methods with data fusion,a quantitative analysis model for melamine was developed by combining partial least squares regression (PLSR) with a data fusion method. The results showed that the prediction accuracy of the absorption spectra improved significantly after low-level data fusion. After medium-level data fusion,the prediction precision of the competitive adaptive reweighted sampling (CARS) method was found to be notably better than that of the successive projection algorithm (SPA). High-level data fusion yielded the highest prediction accuracy,with a prediction correlation coefficient (Rp) of 0.99982 and a root mean square error of prediction (RSMEP) of 0.14％. This method provided a rapid,non-destructive,and accurate technique for melamine quantification in milk powder,offering a promising new tool for food additive quantitative analysis.
Terahertz time-domain spectroscopypartial least squares regressiondata fusionquantitative analysis modelmelamine
国家科技期刊平台
NSTL
万方数据
