Hyperspectral Scattering Profiles for Prediction of Beef Tenderness

Tenderness is one of the most important factors affecting consumer perception of eating quality of meat. In this study, 21 crossbred steers between 11 and 13 months old and with live weights ranging from 100 to 140kg were slaughtered in a commercial plant. Beef samples of strip loin cut from 21 crossbred steers postmortem for 48h were aged for 5 days and analyzed their tenderness, i.e. Warner-Bratzler shear force (WBSF). A laboratory hyperspectral imaging system (400-1100 nm) was developed for assessing beef tenderness. Spectral characteristics of hyper-spectral images were extracted to establish prediction model of beef tenderness. The wavelength 772 nm had the highest correlation with beef WBSF among all single wavelengths. Four feature bands, 772 nm, 680 nm, 533 nm and 485 nm were selected for predicting WBSF. Multi-linear regression (MLR) was performed using for calibration samples. The prediction model gave good prediction values of beef WBSF with the correlation coefficient (r) = 0.94 and the standard error of prediction (SEP) of 1.21kg/cm2. This research demonstrated that the hyperspectral imaging technique is useful for nondestructive determination of beef tenderness.