This study proposes a deep learning-based real-time variable flow control system using the segmentation of fruit trees in a pear orchard. The real-time flow rate control, undesired pressure fluctuation and theoretical modeling may differ from those in the real world. Therefore, two types of preliminary experiments were conducted to examine the linear relationship of the flow rate modeling. Through preliminary experiments, the parameters of the pulse width modulation (PWM) controller were optimized, and a field experiment was conducted to confirm the performance of the variable flow rate control system. The field test was conducted for three cases: all open, on/off control, and variable flow rate control, showing results of 56.15 (\(\pm 17.24\))%, 68.95 (\(\pm 21.12)\)% and 57.33 (\(\pm 21.73\))% for each control. The result revealed that the proposed system performed satisfactorily, showing that pesticide use and the risk of pesticide exposure could be reduced.

本研究提出了一种基于深度学习的实时可变流量控制系统，该系统使用梨园中果树的分割。实时流量控制、不希望的压力波动和理论建模可能与现实世界中的不同。因此，进行了两种类型的初步实验来检验流量建模的线性关系。通过初步实验，优化了脉宽调制（PWM）控制器的参数，并进行了现场实验，以确认变流量控制系统的性能。现场测试了三种情况：全开、开/关控制、变流量控制，结果分别为56.15( \(\pm 17.24\) )%、68.95( \(\pm 21.12)\) %和57.33 (\(\pm 21.73\) )% 用于每个控件。结果表明，所提出的系统表现令人满意，表明可以减少农药的使用和农药暴露的风险。