Authentication of rice using near-infrared reflectance spectroscopy

The authentication of rice (Korean domestic rice vs. foreign rice) has been attempted using near-infrared spectroscopy (NIRS). Two sample sets (n(1) = 280 and n(2) = 200) were used to obtain calibration equations and the spectral regions used for this study were 500-600 nm, 700-900nm, and 980-2,498 nm. Modified partial least square (MPLS) regression was used to develop the prediction model. The standard error of cross validation (SECV) and the r(2) were 0.165 and 0.91 respectively for 1st calibration set and 0.165 and 0.93 for 2nd calibration set respectively. The results of the independent validation (n(3) = 80) showed that all of 80 samples were identified correctly. Even though authentication of rice was performed successfully using NIRS, the calibration statistics in this study showed that further effort is needed for implementation of NIRS for authentication of rice for industry purposes.