Design and Development of a Web Application for Potato Leaf Diseases Detection Using CNN

Potato is a major crop which is produced worldwide and to ensure its fine yield, timely and accurate detection of possible disease is necessary for implementing effective measures. The current methods involve looking at the crops manually before harvesting which gives various challenges such as human errors and other models may lead to loss of specialized feature extraction. Therefore, this study is an innovative approach which involves a CNN pre-trained model to detect plant disease in potato leaves and provide farmers with a web-based application to help prevent its damage and reduce crop losses. Convolution Neural Networks (CNN) have shown a promising step to classify plant images based on the disease type. The dataset from Plant Village has healthy potato leaf images and images of leaf disease with Early and Late Blight. The proposed methodology gives 99.77% accuracy with 30 epochs for classifying leaf diseases to be used for finding appropriate solutions.