基于随机森林模型对猕猴桃喷雾效果影响的喷雾机参数研究

Impact of Sprayer Parameters on Kiwi Fruit Spraying Effect Based on Random Forest Model

在猕猴桃避雨棚内进行喷雾试验,测定不同行驶速度、喷嘴直径和喷雾压力的作业参数下,不同高度叶片的雾滴密度、雾滴粒径、雾滴覆盖率,采用随机森林模型工具对喷雾效果进行分析。结果表明:机具运行参数以及对喷雾效果3个指标的平均重要性排序为喷嘴直径>喷雾压力>作业速度;在机具作业时,喷嘴直径越小喷雾压力越大,在机具作业速度越慢的情况下,雾滴密度越大雾滴覆盖率也越高。建立基于随机森林模型的喷雾效果预测模型,对不同高度的猕猴桃叶片喷雾效果试验数据进行预测,雾滴密度真实值与预测值R^(2)达0.8759,雾滴覆盖率真实值与预测值R^(2)达0.8208,预测效果较好。探讨遥控履带自走式风送喷雾机作业参数对喷雾效果的影响,以期为提高机具作业质量和果农的使用效率提供技术参考。

The spray test was conducted in the kiwifruit shelter to determine the fog droplet density,size and coverage rate of leaves at different heights under different operating parameters such as working speed,nozzle diameter and spray pressure.The random forest algorithm model was used to analyze the spray effect.The results showed that the average importance of the operating parameters on the density,particle size and coverage rate of the droplets was as follows:nozzle diameter>spray pressure>operation speed.The smaller the nozzle diameter,the larger the spray pressure and the slower the machine operation speed,the higher the droplet density and the higher the droplet coverage rate.The random forest prediction model of spray effect was established to predict the experimental data of kiwifruit leaf spray effect at different heights.The real and predicted values of droplet density and droplet coverage R^(2) reached 0.8759,and the predicted values of droplet coverage R^(2) reached 0.8208,indicating a good prediction effect.The influence of operating parameters of remote control track self-propelled air-fed sprayer on spraying effect was discussed,in order to provide technical reference for improving the working quality of the machine and the efficiency of fruit growers.