VISIBLE AND NEAR-INFRARED REFLECTANCE SPECTROSCOPY FOR DETERMINING PHYSICOCHEMICAL PROPERTIES OF RICE

ABSTRACT. To assess rice grain quality, various types of analyses of the physicochemical properties of rice must be conducted. Although several methods exist, different properties require specific instruments and chemicals, and most of them are time-consuming. The authors have therefore developed calibration models for determining multiple physicochemical properties of rice using a single instrument, a visible and near-infrared (VIS/NIR) reflectance spectrometer. These models were used to measure the physicochemical properties of 61 samples of Japanese short-grain, non-waxy rice. The results of partial least squares (PLS) regression modeling with full cross-validation methods indicated that reasonably accurate models (correlation coefficient [r] of the validation greater than 0.8) could be obtained for moisture content, sound whole kernel, and appearance (whiteness, translucency, and color) of brown rice, and for moisture content, appearance, and amylogram characteristics (maximum viscosity and breakdown) of milled rice. The level of accuracy was sufficient for classifying rice samples into qualitative groups. Thus, the results demonstrate that VIS/NIR spectroscopy can be used to determine the physicochemical properties of rice and assess its quality