Research on Tea Shoot Monitoring and Recognition Based on YOLOv5 Improved Algorithm

Tea as an important economic crop in China, the national demand in the growing, in order to meet the growing demand, this paper through in-depth research algorithms to form a solution to the characteristics of tea shoots, such as colour and shape. Although existing recognition models can distinguish the basic features of tea shoots, their recognition ability is often limited in complex real-world scenarios. Therefore, to address this aspect, this paper first proposes how to accurately identify the quantity and judge the quality of tea shoots in complex scenes. While the problems of light, object coupling and occlusion are the research features of this paper, the neural network with YOLOv5 enhanced algorithmic model is mainly used to carry out the research of real-time monitoring of tea shoots. A complete software and hardware system of real-time monitoring system was constructed in the field environment by highly integrated WLAN SoC chip. Through the software design, the real-time monitoring of tea shoots is achieved, and the acceleration and displacement functions of the monitoring system are carefully debugged and optimised, and the results show that the system accurately identifies and monitors the tea shoots in a variety of complex field environments, which provides powerful technical support for tea yield prediction and pest control, and has high application value.