Rapid and Nondestructive Determination of Glucose in Meat by Near Infrared Spectroscopy Model

Glucose content is one of crucial factors that affecting the nutritional quality, taste and flavor of meat. As the classical determination method need complex and destructive sample preparation, laborious and tedious time, near infrared spectroscopy could be a reliable method in a rapid, nondestructive and convenient method. Therefore, the study mentioned by the title was proposed. Meat tissue from the hypoglycemia group, normal group, and hyperglycemia group was recorded as research object, and the spectral pre-treatment methods, spectral ranges, and latent variables were screened and optimized to build a near infrared spectral model for determination of glucose in meat tissue. The optimal near infrared spectral model was built with vector normalization for spectral pre-treatment methods, spectral range in 4 246-9 403 cm^-1, and 17 of latent variables. The correlation coefficients and residual predictive deviations of the calibration set and validation set were 0.990 3 and 0.975 6, as well as 7.23 and 4.54, respectively. The root mean square errors of cross-validation and prediction were 0.057 0 and 0.061 4, respectively.