摘要
快速准确地识别黄瓜病害类型是病害防治的前提,针对黄瓜病害图像识别准确度不高的问题,提出一种基于AI图像处理技术的卷积神经网络模型方法,以便提高病害图像识别率。首先,利用AI图像处理技术对病害图像进行预处理;其次,选定卷积神经网络模型及参数,利用样本数进行训练并加载到Web程序端;最后,在四种常见黄瓜病害图像构建的数据库上进行测试,识别准确率均大于90%。结果表明,基于AI图像处理技术的卷积神经网络模型能够准确识别病害种类,为田间开放环境下实现病害的快速识别提供了依据。
Rapid and accurate identification of cucumber disease types is the premise of disease control.In view of the low accuracy of cucumber disease image recognition,a convolution neural network model method based on AI image processing technology is proposed to improve the disease image recognition rate.Firstly,AI image processing technology is used to preprocess the disease image;Secondly,the convolutional neural network model and parameters are selected,trained by the number of samples and loaded into the web program;Finally,the accuracy of recognition is more than 90%in the database of four common cucumber disease images.The results show that the convolution neural network model based on AI image processing technology can accurately identify the types of diseases,which provides a basis for the rapid identification of diseases in the open field environment.
作者
刘小红
Liu Xiaohong(Software Engineering Department of Hunan University of Information Technology,Changsha 410151)
出处
《现代计算机》
2022年第19期71-74,共4页
Modern Computer
基金
湖南省教育厅科学研究项目(20C1311)。
关键词
病害识别
AI
深度学习
卷积神经网络
disease identification
AI
deep learning
convolutional neural network