Potato Leaf Disease and its Classification Using Deep Learning

Abstract

Potatoes are one of the most extensively consumed foods item, ranking as the 3rd largest staple food consumed throughout the world. Also, the demand for potato is expanding dramatically in the market, particularly due to the worldwide Coronavirus pandemic. However, potato diseases are the major cause of loss in the quality and quantity of the yield. Potato leaf blight is one of the most damaging global plant diseases because it impairs the productivity and quality of potato crop and badly impacts both individual farmers and the agricultural economy. Inappropriate classification and late diagnosis of the disease's type will severely impair the state of the potato plant. This study describes an architecture developed for potato leaf blight classification. This design depends on Deep Convolutional Neural Network (CNN). The methodology also takes use of Data Augmentation. The training dataset is visibly separated into three categories, namely, healthy leaves, early blight leaves and late blight leaves. The number of photos in the collection is 3000. The proposed design achieved an overall mean testing accuracy of 98%. The testing accuracy of the proposed approach was compared with that of comparable works, and the proposed architecture achieved improved accuracy compared to the related works.