CNN-Based Identification of Rice Leaf Diseases

The identification of rice leaf diseases through visual inspection remains a challenge due to the need for expert knowledge and the high similarity among various disease symptoms. This study aims to develop an automatic classification system using a Convolutional Neural Network (CNN) to identify rice leaf diseases based on leaf images. The dataset comprises 15,030 images distributed across nine classes: eight diseases and one healthy category. The data were split into training, validation, and testing subsets with an 80:10:10 ratio. The CNN model was built using TensorFlow and Keras on the Google Colab platform and trained for 13 epochs with image normalization and data augmentation techniques to improve generalization. The model achieved a training accuracy of 89.30%, a validation accuracy of 85.64%, and a test accuracy of 84.52%. These results demonstrate that the CNN model can effectively classify rice leaf images with high precision, highlighting its potential application in intelligent agricultural diagnostics.