Computer vision-based algorithms, especially Convolutional Neural Networks (CNN) and You Only Look Once (YOLO), have become the leading approaches in plant disease detection. CNN excels in extracting complex visual features for disease classification, while YOLO provides high-efficiency real-time object detection capabilities. Both algorithms have shown promising results in various studies, especially with controlled datasets. However, challenges remain in their application in real-world conditions, such as environmental diversity, overlapping symptoms, and poorly annotated data. Future research has the potential to optimize these algorithms through the development of lighter models, the use of transfer learning techniques, and multi-modal data integration. In addition, further exploration of a wider range of diseases, crops, and environmental conditions can expand the application of these algorithms. By leveraging these innovations, computer vision-based plant disease management can be improved to support sustainable precision agriculture.
