Spot Disease Detection in Banana Leaves Using Advanced Machine Learning Techniques

Banana crops are vital for food security in tropical regions but are threatened by diseases like banana leaf spot disease (LSD), which can severely impact yields. Traditional identification methods are labor-intensive and error-prone, while existing automated systems often lack sufficient accuracy in diverse environments, revealing a significant gap in disease management. This study develops a convolutional neural network (CNN) model based on the ResNet50 architecture to accurately classify banana leaf diseases into four categories: Cordana, Sigatoka, Healthy, and Pestalotiopsis. Using a dataset from Kaggle, the model employs data augmentation to simulate varied field conditions, improving its generalization capabilities. Preliminary results indicate that the ResNet50-based model achieves promising accuracy in disease classification, suggesting its potential as a valuable tool for early detection. This automated solution could be integrated into mobile devices, facilitating proactive disease management and supporting sustainable banana farming practices.