Abstract:
Deep Learning (DL) has emerged as a powerful technology in modern agriculture,
revolutionizing practices like precision farming and disease management.
Traditional methods for detecting diseases in bean leaves are manual,
timeconsuming, and require domain expertise, posing challenges in large-scale
operations. Convolutional Neural Networks (cnns), supported by techniques like
Transfer Learning (TL) and ensemble modeling, provide an automated, efficient, and
scalable solution for disease classification. This research evaluates and compares the
performance of state-of-the-art DL models, including VGG-19, resnet50,
mobilenetv2, Vision Transformer (vit), and Xception, to classify bean leaf diseases
effectively. VGG-19 achieved the highest accuracy and was deployed as a web-based
application for real-time disease detection. This study demonstrates how integrating
DL into agricultural workflows can enhance productivity, promote sustainability,
and ensure global food security by offering a precise and scalable solution to identify
and manage bean leaf diseases
