Detecting γ-aminobutyric acid and folates in wheat seedlings using hyperspectral imaging

Abstract

This study introduces the application of hyperspectral imaging (HI) technology for the non-destructive detection of γ-aminobutyric acid (GABA) and folates, which are bioactive compounds in wheat seedlings. Analyzing the GABA and folate content in 210 wheat varieties distributed across China provides an ideal dataset for model prediction. By comparing various preprocessing, filtering, and modeling techniques, predictive models with high accuracy were established. The 1st-GA-PLS model effectively predicts GABA content (RMSEC = 27.887, Rc² = 0.803, RMSEP = 25.640, Rp² = 0.858, RPD = 2.576), while the S-G-SPA-SVM model achieves very high accuracy in predicting folate content (RMSEC = 9.761, Rc² = 0.937, RMSEP = 18.742, Rp² = 0.894, RPD = 3.065). Our research provides meaningful insights for the intelligent cultivation of wheat seedlings, offering a valuable tool for precision farming and production of wheat seedlings as functional food ingredients.