In recent years, the scarcity of successors for farmers in Japan has emerged as a pressing problem. This issue has drawn the attention of researchers, pushing towards smart agriculture by utilizing robots and other cutting-edge technologies. With cooperation from a local producer, the research team is developing an automated Wood Ear Mushroom harvesting system. In this study, the author conducted an experiment to confirm the extent to which recognition accuracy could be improved by switching the detection algorithm from YOLOv4 to YOLOv7. The results indicated a recognition accuracy improvement of more than 10%.
smart agriculture
automatic harvesting system
deep learning
you only look once