Near-infrared reflectance spectra (1,100 to 2,498 nm) were collected on beef longissimus thoracis steaks for the purpose of establishing the feasibility of predicting meat tenderness by spectroscopy. Partial least squares (PLS) analysis (up to 20 factors) and multiple linear regression (MLR) were used to predict cooked longissimus Warner-Bratzler shear (WBS) force values from spectra of steaks from 119 beef carcasses. Modeling used the combination of log(1/R) and its second derivative. Overall, absorption was higher for extremely tough steaks than for tender steaks. This was particularly true at wavelengths between 1,100 and 1,350 nm. For PLS regression, optimal model conditions (R2 = .67; SEC = 1.2 kg) occurred with six PLS factors. When the PLS model was tested against the validation subset, similar performance was obtained (R2 = .63; SEP = 1.3 kg) and bias was small (<.3 kg). Among the 39 samples in the validation data set, 48.7, 87.7, and 97.4% of the samples were predicted within 1.0, 2.0, and 3.0 kg, respectively, of the observed Warner-Bratzler shear force value. The optimal PLS model was able to predict whether a steak would have a Warner-Bratzler shear force value < 6 kg with 75% accuracy. The R2 of MLR model was .67, and 89% of samples were correctly classified (< 6 vs > 6 kg) for Warner-Bratzler shear force. These data indicate that NIR is capable of predicting Warner-Bratzler shear force values of longissimus steaks. Refinement of this technique may allow nondestructive measurement of beef longissimus at the processing plant level.

Near-infrared (NIR) spectroscopy was used to measure the amount of salt (NaCl) in canned cured hams. Calibration with 20 samples produced a high correlation (r = 0.96) between salt contents determined by chemical analysis and by NIR using the second derivative of log (1/reflectance) values at 1806 nm. Salt contents of nineteen unknown samples were predicted with a standard error of prediction of 0.17% NaCl. Salted beef, salted fresh ham (uncured), and salted water model systems demonstrated that the ability to measure salt by NIR is due to the shift in the water spectrum caused by salt-induced changes in the amount of hydrogen bonding.

Near infrared (log 1/reflectance) spectra of samples of ground wheat were approximated by a linear combination of spectra of known constituents, the approximation satisfying a least-squares criterion. The appropriate coefficients of these linear combinations were then linearly correlated with the protein and moisture content of the samples. Two extensions of previously reported curve-fitting techniques were made. First, multilinearly correlating several of the curve fit coefficients with the chemical data improved the standard errors. Second, using sample spectra as components, rather than pure constituent spectra, improved the standard errors to a point where they became comparable to those obtained by currently used derivative methods. The samples covered a protein range of 10 to 19%. Correlation coefficients reached 0.998 for protein, corresponding to a standard error of prediction of 0.15%. Parameters examined included spectral region, smoothing, and wavelength shifting. Results with reflectance spectra of sample sets with large particle size variation and high noise are also reported.

The R 2 was 0.612 when the spectra between 1100 nm and 1350 nm were analyzed. When the second derivatives of the spectral data were used, the R 2 of the PCR model to predict WB shear force of the meat was 0.633 for the full spectral range of 1100 to 2498 nm and 0.616 for the spectral range of 1100 to 1350 nm.

Cereal Chem. 77(1):86-88