<![CDATA[Rice (Oryza sativa) is a strategic commodity for food security in Indonesia, yet it is highly vulnerable to diseases such as bacterial leaf blight (blight), blast, and tungro, which can significantly reduce productivity. Early detection of these diseases through manual observation by farmers is often inaccurate and slow. This study aims to implement the YOLOv11 algorithm, a deep learning-based approach, to detect rice plant diseases from leaf images with high accuracy. The research method follows the CRISP-DM (Cross Industry Standard Process for Data Mining) framework, encompassing business understanding, data collection, data preparation, modeling, and evaluation. The dataset consists of 500 rice leaf images classified into three disease categories. The data was processed through augmentation and resizing to balance class distribution and standardize image dimensions. The YOLOv11 model was trained with parameters set at 100 epochs, an image size of 224x224 pixels, and a batch size of 32. Evaluation results demonstrate that the model achieved 95% accuracy, with average precision and recall exceeding 95%. The confusion matrix revealed excellent classification performance, particularly for tungro disease (100% accuracy). The model also proved efficient in prediction, with an inference time of 8.2 milliseconds per image. In conclusion, this research confirms the effectiveness of YOLOv11 for rice disease detection based on leaf images. Recommendations for future development include expanding dataset diversity, integrating the model into mobile applications, and conducting field tests to validate real-world performance. Keywords: YOLOv11, rice disease detection, deep learning, leaf image, computer vision.]]>
