Detection and Classification of Potato Leaves Diseases Using Convolutional Neural Network and Adam Optimizer

This study explores the development of a Convolutional Neural Network (CNN) with Adam Optimizer for detecting potato leaf blight, a critical concern in agriculture due to the economic importance of potatoes. The research focuses on developing a highly accurate model to distinguish between healthy and diseased potato leaves using image data. Key steps include dataset collection, image preprocessing with scaling, augmentation, and normalization, and the construction of a CNN architecture featuring convolutional, max-pooling, dense layers, and dropout regularization. The model’s performance, evaluated through accuracy, precision, recall, and F1-scores derived from confusion matrix analysis, achieved an impressive 96.88% accuracy. Specific detection metrics include 0.76 precision, 0.93 recall, and 0.84 F1-score for healthy leaves, and near-perfect scores for Late and Early Blight detection. The findings underscore the efficacy of CNNs combined with Adam optimization in agricultural disease diagnosis, offering a reliable tool for early potato leaf blight detection.