AI Summary & Key Takeaways
AI InsightThese results showed that the NIR-based models provided a distinct advantage over MIR-ATR-based models in accurately estimating the ratio of the dried green pea seeds in binary mixtures.
📌 3 Key Takeaways
Abstract
Near-infrared (NIR) diffuse reflectance and Mid-infrared-attenuated total reflectance (MIR-ATR) spectroscopy were evaluated to determine the dried green pea seed adulteration in ground pistachio nuts. Sixty-three samples (51 for calibration and 12 for external validation sets) of ground pistachio nuts were deliberately adulterated with varying levels of dried green pea seeds (ranging from 0 to 50% (w/w)). Subsequently, both NIR and MIR-ATR spectra of the samples were collected separately. The quantitative predictions of green pea ratio in the samples were achieved using PLSR (Partial Least Squares Regression). Based on the PLSR models, SEP (Standard error of prediction) values of the models were 2.55 and 9.14% for NIR and MIR-ATR, respectively. The rPred (Correlation coefficient of prediction) values of the models were 0.99 and 0.80 for NIR and MIR-ATR, respectively. Additionally, the nondimensional values of RPD (Residual Predictive Deviation) for NIR and MIR-ATR spectra were calculated as 5.7 and 1.6, respectively. These results showed that the NIR-based models provided a distinct advantage over MIR-ATR-based models in accurately estimating the ratio of the dried green pea seeds in binary mixtures. Therefore, NIR spectroscopy has the potential and could be implemented in the routine applications of green pea detection in ground pistachio nuts.