Predicting micronutrients of wheat using hyperspectral imaging.

Abstract Micronutrients are the key factors to evaluate the nutritional quality of wheat. However, measuring micronutrients is time-consuming and expensive. In this study, the potential of hyperspectral imaging for predicting wheat micronutrient content was investigated. The spectral reflectance of wheat kernels and flour was acquired in the visible and near-infrared range (VIS-NIR, 375–1050 nm). Afterwards, wheat micronutrient contents were measured and their associations with the spectra were modeled. Results showed that the models based on the spectral reflectance of wheat kernel achieved good predictions for Ca, Mg, Mo and Zn (r2 > 0.70). The models based on the spectra reflectance of wheat flour showed good predictive capabilities for Mg, Mo and Zn (r2 > 0.60). The prediction accuracy was higher for wheat kernels than for the flour. This study showed the feasibility of hyperspectral imaging as a non-invasive, non-destructive tool to predict micronutrients of wheat.