摘要
农作物病害分类是细粒度图像分类的一个热门领域。文章采用一种基于Co-Location的细粒度农作物病害分类方法,在模拟真实场景的MultiplePlant测试集下进行农作物病害的研究,比较各个网络在得到的模型的准确率结果为98.36%,证明该方法充分保留了关键病害位置的特征信息,并且一定程度上能够抵抗噪声干扰。
Crop disease classification is a hot field of fine-grained image classification.In this paper,a fine-grained crop disease classification method based on Co-location is used to study crop diseases under the multipleplant test set simulating the real scene.Compared with each network,the accuracy of the model is 98.36%,which proves that this method fully retains the characteristic information of key disease locations and can resist noise interference to a certain extent.
作者
徐妍
Xu Yan(Zhejiang Agriculture and Forestry University,Hangzhou 310000,Zhejiang,China)
出处
《农业技术与装备》
2021年第10期97-98,共2页
Agricultural Technology & Equipment
关键词
图像分类
农作物病害识别
协同定位
细粒度分类
image classification
identification of crop diseases
collaborative positioning
fine grained classification
