The rapid detection of the tannin content of grains based on hyperspectral imaging technology and chemometrics

Tannins are one of the key components that constitute the nutritional structure of grains, and have a significant impact on the quality of grain products. The detection of tannins by chemical methods is both time- and manpower-consuming. In this study, hyperspectral imaging (HSI) technology was combined with an optimized algorithm, iteratively variable subset optimization (IVSO), in conjunction with the variable importance in projection (VIP) method to rapidly detect the tannin content of grains (sorghum, rice, glutinous rice, wheat, and corn). The IVSO algorithm and VIP method were employed to extract characteristic wavelengths. Subsequently, three tannin content prediction models based on full and characteristic wavelengths (BPNN, PSO-LSSVM, and DF) were respectively established. A comparative analysis demonstrated that the DF model based on the characteristic wavelengths had the best effect (R-p(2) = 0.9922 g/100 g, RPD = 11.3), and the tannin content of grain particles was visually analyzed by using the best prediction model. The results show that HSI can realize the rapid non-destructive detection of the tannin content of grain particles, which provides a new technical reference for the food industry to improve product quality.