Abstract: The volatile organic compounds of wheat with different mildew degrees were detected using gas chromatography-mass spectrometry (GS-MS). Principal components analysis (PCA) of the volatile organic compounds was performed to effectively distinguish the different degrees of moldy wheat in order to provide an experimental basis for visual identification of moldy wheat. Olfactory visualization as a rapid and convenient method was employed to detect fresh and moldy wheat with different mildew degrees, and the data obtained were processed by PCA followed by linear discriminant analysis (LDA) and K-nearest neighbor (KNN) algorithm. The KNN and LDA models gave a recognition rate of 95.83% and 85.40%, respectively. These results conclusively show that olfactory visualization technology can allow fast, non-destructive and accurate detection of moldy wheat.

Key words: moldy wheat, volatile organic compounds, olfactory visualization