Implementasi Convolutional Neural Network (CNN) dalam Diagnosa Penyakit Daun Padi Berdasarkan Citra Digital

This study investigates the implementation of Convolutional Neural Network (CNN) in classifying rice leaf diseases based on digital images. The model classifies three types of diseases: Bacterial Leaf Blight, Rice Blast, and Rice Tungro Virus. A dataset of 240 images was obtained from Kaggle, with 80 images per class. Four training scenarios were applied using 25, 50, 75, and 100 epochs. Preprocessing steps included resizing all images to 150x150 pixels and normalizing pixel values. Evaluation results show that classification accuracy increases with the number of training epochs. The best model was achieved at 100 epochs, yielding a validation accuracy of 91.67% and testing accuracy of 92%. These results demonstrate that CNN is effective in diagnosing rice leaf diseases and can support early detection efforts to strengthen national food security.