Predicting Concentrations of Individual Sugars in Dry Mixtures by Near-Infrared Reflectance Spectroscopy

To support research in nondestructive quality evaluation of fruits and vegetables, near infrared reflectance spectroscopy was used to measure concentrations of fructose, glucose, and sucrose in model systems intended to represent the major constituents of dried apple tissue. Spectra (950-1850 nm) of a series of known samples were recorded, and multiple linear regression techniques were used to relate the concentrations of each sugar to reflectance measurements at selected wavelengths. Reflectance measurements at 3 computer-selected wavelengths for each sugar were used to predict the concentrations in an independent set of samples. Correlation coefficients for actual vs predicted values were 0.995 for fructose, 0.994 for glucose, and 0.986 for sucrose, while the respective standard errors computed as variation from the regression lines were 1.48%, 1.39%, and 1.44%.