Abstract

Quality assurance and guaranteeing meat and eating quality continues to be an ongoing challenge to red meat processors, as many measures of meat quality require resource intensive, destructive or subjective tests. Consequently, there is a need to develop rapid, non-invasive and non-destructive methods for meat quality assessment in commercial processing plants. Therefore, a preliminary investigation was conducted to determine the potential for a hand-held Raman spectroscopic device to predict meat quality and sensory traits measured by an untrained consumer panel. To this end, 45 beef loins (m. longissimus thoracis) were measured using a 671nm hand-held Raman spectroscopic device. Once spectra were collected, loins were frozen until sensory testing and shear force measurements. Further sub-samples were also collected to measure traits including sarcomere length, particle size, pH and colour. Derived models indicate that using Raman spectra it was possible to predict juiciness and tenderness, with cross validated correlations between the predicted and observed values (R2cv) of 0.42 (RMSEP = 11.29) and 0.47 (RMSEP = 10.52). A tentative band assignment suggests that these predictions may be based on changes to the biochemical characteristics of the muscle asso-ciated with hydrophobicity of proteins, fatty acid composition and the collagen matrix. However, further research is required to determine the repeatability and robustness of these models on a larger independent data set.

SM, Fowler, H Schmidt, and DL Hopkins

Proceedings of the New Zealand Society of Animal Production, Volume 77, Rotorua, 181-184, 2017