Effect of spectral pretreatment on qualitative identification of adulterated bovine colostrum by near-infrared spectroscopy

Identifying the adulterated bovine colostrum is helpful to guarantee the quality of colostrum products and protect the legitimate rights of consumers. Near-infrared (NIR) spectroscopy technology has been proved to be able to predict the main compositions of milk, but the scattering of particles in colostrum has big effect on obtained NIR spectra. To know the effect of spectral pretreatment on qualitative identification of adulterated bovine colostrum by NIR spectroscopy, 154 bovine colostrum samples adulterated with mature milk at the mass concentration of 0–50% were prepared, and their NIR spectra in 833–2500 nm were pretreated using Savitzky-Golay smoothing (S-G), second derivative (SD), multiple scatter correction (MSC), SD after S-G (S-G + SD) and MSC after S-G (S-G + MSC) in this study. The results showed that MSC was the optimum spectral pretreatment method, and the partial least squares discriminant analysis (PLS-DA) model based on full spectra after MSC pretreatment had the highest sensitivity (84.62%), specificity (100%) and accuracy rate (94.74%). Compared with the established PLS-DA model with the spectra without any pretreatment, the accuracy rate was improved by 5.27%. The additional experiment showed that the developed method had good identification performance. This study provides a fast and accurate detection method for qualitative identification of adulterated bovine colostrum.