Near-Infrared Reflectance Determination of Fat, Protein, and Moisture in Fresh Meat

Abstract
Near-infrared spectral absorption properties were determined for emulsified and ground meat samples. Six corrected log numbers associated with the optical responses of 6 filters in the InfraAlyzer were used as multiple independent variables in regression equations. Dependent variables for the equations were moisture, determined by oven-drying; fat, determined by Goldfisch extraction; and protein, determined by the Kjeldahl method. InfraAlyzer log values, replicated 4 times on each emulsified beef sample and 3 times on each ground lamb sample, increased as the samples were warmed from the heat generated in the sample drawer. Variation in temperature of the meat samples was partially responsible for differences in constant terms and in regression coefficients for equations developed on data from different replications of a sample. Multiple correlation coefficients for fat ranged from 0.91 to 0.94 in emulsified meat and from 0.83 to 0.85 in ground meat; for moisture, 0.90-0.94 and 0.83-0.85, respectively; and for protein, 0.80-0.85 and 0.72-0.77, respectively. Overall, near-infrared reflectance shows promise as a rapid method for determining composition of meat. Nevertheless, some aspects of near-infrared reflectance require further attention.

This publication has 2 references indexed in Scilit: