Early Detection of Seasonal Bean Leaf Spot using Deep Convolutional Neural Networks

Identifying early bean leaf spots is essential for minimizing crop loss; however, conventional techniques relying on physical inspection, basic image processing, and feature extraction are subjective, have low precision, and have a narrow range of applications. Conventional techniques also struggle to adjust to a variety of environmental settings and frequently miss early-stage illnesses. To overcome these problems, the proposed system will automatically extract specific features from leaf images using a deep convolutional neural network (CNN). The dataset was a mix, and the trained model underwent preprocessing and augmentation to make it more robust. The evaluation results confirm that the proposed system performs better than the existing systems historically in use, with an accuracy of 96.2%, precision of 95.4%, recall of 94.8%, and F1-score of 95.1%. With an accuracy level of 91.8%, even early-stage detection is good. Confusion matrix analysis does in fact support its high dependability and low rate of incorrect predictions. The scalable approach enhances agricultural outcomes by giving farmers real-time feedback and better disease surveillance.