Application of Vis-NIR and SWIR spectroscopy for the segregation of bison muscles based on their color stability

Abstract

The study objective was to investigate the potential for using visible near-infrared (Vis-NIR) and short wave infrared (SWIR) spectroscopy to segregate bison portions based on muscle types and storage periods. In the Vis-NIR range, the principal component analysis showed clear segregation of the muscles based on storage at retail display d 4 whereas the discrimination based on muscle type was better portrayed in the SWIR region. Furthermore, partial least squares discriminant analysis (PLS-DA) models classified muscles based on muscle type and storage in the Vis-NIR range with the classification accuracy of 97% for calibration and 86% for cross-validation. Finally, the PLS-regression models were developed for the successful prediction of a* value with an R² of 0.88 (RMSEC: 1.57), 0.84 (RMSECV: 1.88), and 0.90 (RMSEP: 1.41), color score with an R² of 0.96 (0.25), 0.95 (0.27), and 0.92 (0.32), and discoloration score with an R² of 0.96 (0.47), 0.93 (0.63), and 0.93 (0.56) for calibration, cross-validation, and prediction, respectively.

Keywords: Classification; Meat color; Partial least squares discriminant analysis; Principal component analysis; Short wave infrared; Visible near-infrared.