Deep Learning-based Paddy Doctor for Sustainable Agriculture

In promoting sustainable agriculture, the prompt and precise identification of plant diseases and pests becomes critical in mitigating economic setbacks via timely interventions. Utilizing EfficientNet, a deep learning-based convolutional neural network (CNN), we categorized images using open-source datasets on various rice strains. Our empirical findings confirm that our suggested model architecture attains a commendable accuracy of 91% with imbalanced classes. Discarding blurry images and employing data augmentation techniques enhanced the model's precision in rice disease categorization by an additional 8.4%. The corresponding code will be publicly accessible here.