Monitoring S-ovalbumin content in eggs during storage using portable NIR spectrometer and multivariate analysis

Abstract

The study investigated the feasibility of using portable NIR spectrometer coupled with multivariate analysis to monitor the S-ovalbumin content in eggs during storage. The NIR spectra of egg samples were obtained in the spectral range of 900–1700 nm. Savitzky-Golay (SG) smoothing, wavelet threshold denoising (WTD), multiplicative scatter correction (MSC), standard normal variate (SNV) and their combinations were used to pretreat the raw spectral data. In addition, successive projection algorithm (SPA) and iteratively retains informative variables (IRIV) were employed to extract feature related wavelengths from the full spectrum. Finally, the developed partial least squares regression (PLSR) model based on 14 IRIV feature wavelengths preprocessed by WTD-MSC had the best performance (R2P = 0.86, RMSEP = 9.33%). The result indicates that portable NIR device can be used with caution for nondestructive monitoring of S-ovalbumin content in eggs and has the potential to be applied in the supply chain.