Discrimination of Deoxynivalenol Levels of Barley Kernels Using Hyperspectral Imaging in Tandem with Optimized Convolutional Neural Network

• Published in: Sensors (Basel, Switzerland) Switzerland), 2023-02, Vol.23 (5), p.2668
• Main Authors: Fan, Ke-Jun, Liu, Bo-Yuan, Su, Wen-Hao
• Format: Article
• Language: English
• Subjects: Accuracy; Adaptive sampling; Algorithms; Animal health; Barley; Cereals; Chromatography; Classification; Climate change; Contamination; convolutional neural network; Crop diseases; Data analysis; Deep learning; Deoxynivalenol; feature wavelength selection; Food contamination; Food contamination & poisoning; Hordeum; Humans; Hyperspectral Imaging; Imaging systems; Machine learning; Methods; Neural networks; Neural Networks, Computer; Pesticides; Support Vector Machine; Support vector machines; Wavelet transforms
• ISSN: 1424-8220; 1424-8220
• DOI: 10.3390/s23052668

Deoxynivalenol (DON) in raw and processed grain poses significant risks to human and animal health. In this study, the feasibility of classifying DON levels in different genetic lines of barley kernels was evaluated using hyperspectral imaging (HSI) (382-1030 nm) in tandem with an optimized convolutional neural network (CNN). Machine learning methods including logistic regression, support vector machine, stochastic gradient descent, K nearest neighbors, random forest, and CNN were respectively used to develop the classification models. Spectral preprocessing methods including wavelet transform and max-min normalization helped to enhance the performance of different models. A simplified CNN model showed better performance than other machine learning models. Competitive adaptive reweighted sampling (CARS) in combination with successive projections algorithm (SPA) was applied to select the best set of characteristic wavelengths. Based on seven wavelengths selected, the optimized CARS-SPA-CNN model distinguished barley grains with low levels of DON (