Automatic classification of tomato leaf diseases based on MtConvNeXt model

Tomato leaf diseases significantly threaten global food security and agricultural productivity. Traditional manual diagnosis methods are often subjective and time-consuming, prompting the need for efficient automated systems. This study introduces MtConvNeXt, an enhanced ConvNeXt model incorporating a ternary parallel attention mechanism and an SVM classifier, for tomato leaf disease classification. Trained on a dataset of 11,000 images across ten disease categories, MtConvNeXt achieves 96.15% recognition accuracy, outperforming the original ConvNeXt model. Ablation studies and confusion matrix analysis confirm the enhancements’ effectiveness. This research advances automated disease diagnosis in agriculture, potentially improving crop yield and quality while minimizing the need for expert consultations.