A Deep Learning-Based System for Monitoring the Number and Height Growth Rates of Moso Bamboo Shoots

The number and growth of new shoots are very important information for bamboo forest cultivation management. At present, there is no real-time, efficient accurate monitoring method. In this study, a fixed webcam was applied image capture, optimized YOLOv4 used to model the detection moso shoots, strategy sorting screening proposed track each shoot. change in height obtained according boxes. experimental results show that system can remotely automatically obtain pixel shoot at any given time. average relative error variance were 1.28% 0.016%, respectively, those corresponding −0.39% 0.02%. This be series purposes, such as daily or weekly rate stations trends selected shoots.