Automated disease detection in banana using convolutional neural networks (CNNs)

Banana is crucial to global food security and the agricultural economy, but is very susceptible to many plant diseases that reduce yield and quality. Early and accurate disease detection is key to mitigating losses and better crop management. In this work, we propose an automated disease detection system for banana plants using Convolutional Neural Networks (CNNs). A carefully selected dataset of pictures of banana leaves was used to build a deep learning model for frequent illnesses like Bacterial Wilt, Fusarium Wilt, and Black Sigatoka. To increase detection accuracy, data augmentation and hyperparameter tuning were used to refine the suggested CNN architecture. Experimental results show that our model achieved 98.5% accuracy on the test dataset beating traditional machine learning classifiers and pre-trained models in terms of precision, recall and F1-score. The proposed system is real- time, cost-effective and scalable for banana disease diagnosis, enabling early intervention and reducing manual inspection dependency. In the future, the model will be expanded to make it more generalizable across various climatic circumstances and integrated with edge computing devices for real-time field deployment.