MobileNet-Based Real-Time Identification of Tomato Leaf Pathologies for Crop Health Management

This study introduces a MobileNet-based deep learning (DL) model for the organization of tomato leaf illnesses, to enhance automated disease detection in agricultural environments. The methodology is designed to categorise tomato leaves into four classifications: Scab, Rot, Alternaria, and Healthy. The MobileNet architecture’s lightweight and economical design guarantees excellent accuracy while minimizing computing requirements, rendering it appropriate for real-time applications in resource-constrained settings. To improve generalization and handle variability in leaf appearance, various data augmentation techniques were employed. The suggested MobileNet model achieves a remarkable accuracy of 96.02% on the tomato dataset, outperforming traditional CNN models and ensemble approaches. Future research could focus on integrating the model with IoT devices for real time monitoring and disease detection in the field and expanding the dataset with a diverse set of images. This would enable a comprehensive solution for crop health management, supporting farmers in optimizing crop yield and mitigating disease loss.