摘要
虫害是影响树木健康的重要因素,虫害防治效果取决于虫害预报的及时性和准确性。而传统方法通过人工观察(捕捉)来预报虫害爆发,具有监测预报效率低等问题。因此本文利用YOLOv5算法建立病虫害监测预报系统,快速采集实时病虫害照片,利用Focus结构的切片操作对图像进行切片,经过卷积操作后得到新的图像,并将深度学习模型与YOLOv5算法模型进行对比。实验结果表明,基于YOLOv5算法的监测系统可以有效识别叶枯病,识别准确率较高,而深度学习模型的褐斑病识别效果较差。且害虫预警模型的准确率为0.98,召回率为0.97,F1值为0.97,可满足实际林地病害虫监测。
Insect pests are important factors affecting the health of trees,and the effectiveness of pest control depends on the timeliness and accuracy of pest forecasting.The traditional method of forecasting pest outbreaks through manual observation(capture)has the problem of low monitoring and forecasting efficiency.Therefore,in this paper,YOLOv5 algorithm is used to establish a pest monitoring and forecasting system,which quickly collects real-time pest photos,slices the image using the slicing operation of Focus structure,and obtains a new image after convolution operation,and compares the deep learning model with the YOLOv5 algorithm model.The experimental results show that the monitoring system based on the YOLOv5 algorithm can effectively recognize leaf blight with high recognition accuracy,while the deep learning model has poor recognition of brown spot disease.And the accuracy rate of the pest warning model is 0.98,the recall rate is 0.97,and the F1 value is 0.97,which can satisfy the monitoring of actual woodland diseases and pests.
作者
张娟
ZHANG Juan(Lanzhou Ecological Forestry Experimental Site,Lanzhou 730085,Gansu,China)
出处
《林业科技情报》
2025年第3期61-63,共3页
Forestry Science and Technology Information
关键词
病虫害
监测预报系统
YOLOv5算法
病虫害照片
Pests and diseases
monitoring and forecasting system
YOLOv5 algorithm
pest and disease photos
