This study evaluated a portable multispectral imaging (MSI) system for simultaneously assessing chicken meat quality, including freshness and authenticity detection. For freshness, total aerobic counts and MSI analyses were performed on fresh and thawed samples throughout storage at 4 °C. For authenticity (product condition and origin), Greek and Danish chicken samples, both fresh and thawed, were analyzed in separate laboratories using identical instruments. Data were modeled using PLS-R, kNN, and SVM. Model performance for total viable count prediction was evaluated via R2 and RMSE, while classification used accuracy, specificity, recall and precision. PLS-R beta coefficients highlighted the contribution of specific wavelengths. For Greek chicken fillets, kNN achieved the best performance on fresh samples (RMSE = 0.347, R2 = 0.979), while PLS-R performed best on thawed samples (RMSE = 0.787, R2 = 0.859). Wavelength 460 nm was the most important for all freshness predictions. Differences between Danish and Greek samples were observed in classification performance, optimal algorithms and key wavelengths. For origin classification (using fresh and thawed samples), models reached near-perfect accuracy, with PLS-DA highlighting 660 nm and 850 nm as most significant. These results demonstrate the MSI system’s potential for the rapid, accurate and simultaneous evaluation of multiple chicken meat quality attributes using a single instrument.
Lytou et al. (Mon,) studied this question.