Deoxynivalenol Detection beyond the Limit in Wheat Flour Based on the Fluorescence Hyperspectral Imaging Technique

Area of Science:

• Food Science
• Analytical Chemistry
• Spectroscopy

Background:

• Deoxynivalenol (DON) contamination in wheat flour is a significant global food safety issue.
• Accurate detection methods are crucial for monitoring DON levels and preventing health risks.

Purpose of the Study:

• To develop and validate a method for detecting excessive Deoxynivalenol (DON) content in wheat flour.
• To evaluate the effectiveness of fluorescence hyperspectral imaging (FHSI) combined with machine learning for DON detection.

Main Methods:

• Wheat flour samples were artificially contaminated with varying DON concentrations.
• Fluorescence hyperspectral images were collected and preprocessed using SG smoothing and normalization.
• Feature band selection was performed using algorithms like CARS and SPA.
• Classification models, including Random Forest (RF), Support Vector Machine (SVM), and Convolutional Neural Network (CNN), were employed.

Main Results:

• The SG-CARS-RF and SG-CARS-SVM models achieved high performance, with recall rates up to 98.95% and accuracy up to 97.78%.
• The Random Forest algorithm demonstrated robust performance, further validated by ROC curves.
• The CNN model achieved a recognition accuracy of 97.78% for the test set.

Conclusions:

• Fluorescence hyperspectral imaging (FHSI) is a feasible and promising technique for detecting DON contamination in wheat flour.
• Machine learning algorithms, particularly RF and SVM, effectively discriminate between acceptable and excessive DON levels.
• This approach offers potential for rapid and accurate food safety assessments regarding DON in wheat products.