Wiederhold, William (2011-05). Visible-NIR, Electrical Impedance, pH, and CIE L*, a*, and b* Color Space Values to Predict Beef Tenderness. Master's Thesis.
Thesis
Predicting tenderness in today's beef supply could be advantageous to packers and consumers. In this study (n = 1,137 carcasses), visible-near-infrared, electrical impedance, pH and Minolta CIE L*, a*, and b* color space values were examined as predictors of beef 1, 7, and 14 d Warner-Bratzler (N) or Slice Shear (N) force values as estimators of beef tenderness. Visible-NIR at 350 to 1830 nm, electrical impedance, and color space values were taken at the beef packing plant, along with carcass data. Strip loins were transported to Texas A & M University where pH was taken. Six steaks were taken from the anterior end of the strip loin and randomly assigned to either Warner-Bratzler shear force (WBSF) after 1, 7, or 14 days, or Slice shear force (SSF) after 1, 7, and 14 days of post-harvest aging at 2 degrees C. Shears were taken on assigned days. Shear force values were highly correlated with each other (r = 0.37 to 0.56 for WBSF and r = 0.75 to 0.78 for SSF) (P < 0.05). Within the independent variables, reflectance values for mid-range wavelengths (562nm-1193nm) were found to be most highly correlated with the dependent variables (P < 0.05). pH and color spaces values were more highly correlated (P < 0.05) to slice shears values then to Warner-Bratzler shears force values. Electrical impedance was the least significant with r values of 0.00 to 0.14. When Visble-NIR reflectance values were used in stepwise regression equations to predict 1, 7, or 14 d WBSF or 1, 7, or 14 d SSF, prediction equations for 14 d WBSF and SSF had the highest R^2 (0.14 and 0.36, respectively). Stepwise regression equations that included pH and color space values had the highest R^2 for 7 d WBSF and 1 d SSF (0.22 and 0.28, respectively). Electrical impedance alone in a stepwise regression equation had the highest R^2 for 1 and 14 d WBSF and 1 and 7 d SSF (0.02 and 0.03, respectively). Stepwise regression equations that included pH, color space values, and electrical impedance had the highest R^2 for 7 d WBSF and 14 d SSF (0.25 and 0.24, respectively). When pH, color space values, electrical impedance, and Visible-NIR were used, 7 d WBSF and 1 d SSF had the highest R^2 (0.38 and 0.34, respectively). Stepwise regression equations that included pH, color space values, and Visible-NIR had the highest R^2 for 7 d WBSF and 14 d SSF (0.30 and 0.44, respectively). For predicting 14 d Warner-Bratzler shear force, a R^2 of 0.20 was found using Visible-NIR, pH and color space values. When used, the partial least squares equation predicted tenderness with an 85 percent success rate. For predicting 14 d Slice shear forces, a R^2 of 0.40 was found. When used, the partial least squares equation had a 100 percent success rate of predicting those steaks found tender to be tender for Slice shear force. There was an 85 percent success rate for predicting 14 d Warner-Bratzler shear forces. Both equations still had little to no success in predicting tough steaks. The Visible-NIR can successfully predict tenderness
Predicting tenderness in today's beef supply could be advantageous to packers and consumers. In this study (n = 1,137 carcasses), visible-near-infrared, electrical impedance, pH and Minolta CIE L*, a*, and b* color space values were examined as predictors of beef 1, 7, and 14 d Warner-Bratzler (N) or Slice Shear (N) force values as estimators of beef tenderness. Visible-NIR at 350 to 1830 nm, electrical impedance, and color space values were taken at the beef packing plant, along with carcass data. Strip loins were transported to Texas A & M University where pH was taken. Six steaks were taken from the anterior end of the strip loin and randomly assigned to either Warner-Bratzler shear force (WBSF) after 1, 7, or 14 days, or Slice shear force (SSF) after 1, 7, and 14 days of post-harvest aging at 2 degrees C. Shears were taken on assigned days.
Shear force values were highly correlated with each other (r = 0.37 to 0.56 for WBSF and r = 0.75 to 0.78 for SSF) (P < 0.05). Within the independent variables, reflectance values for mid-range wavelengths (562nm-1193nm) were found to be most highly correlated with the dependent variables (P < 0.05). pH and color spaces values were more highly correlated (P < 0.05) to slice shears values then to Warner-Bratzler shears force values. Electrical impedance was the least significant with r values of 0.00 to 0.14.
When Visble-NIR reflectance values were used in stepwise regression equations to predict 1, 7, or 14 d WBSF or 1, 7, or 14 d SSF, prediction equations for 14 d WBSF and SSF had the highest R^2 (0.14 and 0.36, respectively). Stepwise regression equations that included pH and color space values had the highest R^2 for 7 d WBSF and 1 d SSF (0.22 and 0.28, respectively). Electrical impedance alone in a stepwise regression equation had the highest R^2 for 1 and 14 d WBSF and 1 and 7 d SSF (0.02 and 0.03, respectively). Stepwise regression equations that included pH, color space values, and electrical impedance had the highest R^2 for 7 d WBSF and 14 d SSF (0.25 and 0.24, respectively). When pH, color space values, electrical impedance, and Visible-NIR were used, 7 d WBSF and 1 d SSF had the highest R^2 (0.38 and 0.34, respectively). Stepwise regression equations that included pH, color space values, and Visible-NIR had the highest R^2 for 7 d WBSF and 14 d SSF (0.30 and 0.44, respectively). For predicting 14 d Warner-Bratzler shear force, a R^2 of 0.20 was found using Visible-NIR, pH and color space values. When used, the partial least squares equation predicted tenderness with an 85 percent success rate. For predicting 14 d Slice shear forces, a R^2 of 0.40 was found. When used, the partial least squares equation had a 100 percent success rate of predicting those steaks found tender to be tender for Slice shear force. There was an 85 percent success rate for predicting 14 d Warner-Bratzler shear forces. Both equations still had little to no success in predicting tough steaks. The Visible-NIR can successfully predict tenderness