Pork meat quality classification using Visible/Near-Infrared spectroscopic data

Pork meat is currently classified into five quality groups, based on the combination of three main parameters, namely colour, texture and exudation. The usefulness of Visible/Near-Infrared (VIS/NIR) spectroscopy as a non-destructive method for pork meat quality classification was evaluated. Sixty samples of pork loin from each of the four quality classes of meat were selected and their reflectance spectral data in the range from 350 to 2500 nm with a resolution of 1 nm were obtained. Stepwise regression analysis was used to select the most significant wavebands and this was followed by a discriminant analysis to investigate the ability of the selected wavebands to classify pork meat samples into different categories. Leave-one-out and five-fold cross-validation methods were used to validate the procedure. Pork meat quality classes were correctly classified with 79% accuracy, when a small subset of selected wavebands was used. The results demonstrated the potential application of VIS/NIR spectroscopy in pork meat classification.