Convolutional Neural Network Based Approach for Potato Leaf Disease Detection

Abstract

In India, agriculture accounts for about 17–18% of the Gross Domestic Product (GDP), and potatoes are among the most important and widely recognised staple foods worldwide. Nonetheless, the ongoing threat of potato diseases poses significant challenges to both the quantity and quality of harvests, hindering their increasing importance. Identifying diseases in crop leaves manually is both labor-intensive and inefficient. To tackle these issues, there has been a growing trend toward utilizing advanced technologies, such as image processing, machine learning, computer vision, and deep learning, for the effective diagnosis of plant diseases and pests. The adoption of these automated techniques greatly improves the efficiency of monitoring extensive farms in shorter periods. This study examines one such effective technique: a Convolutional Neural Network (CNN)-based method for detecting diseases like early and late blight on potato plant leaves. The PlantVillage dataset, obtained from Kaggle, was employed in this research, achieving a classification accuracy of 99.61%.