摘 要:葡萄在世界历史上源远流长,科技的发展与人口的流动让葡萄的种植范围不断扩大,位居世界首位,这其中也带来了许多问题,例如葡萄病毒病。提前发现病害可以扭转葡萄产量下降趋势。针对人们肉眼判断准确率低的问题,文章提出了基于 CNN 的葡萄病毒病的图像识别模型,网络包含一个输入层、四个卷积层、四个池化层、两个全连接层和一个输出层,对于文章选取的数据,该模型的精确度达到了 97.25%,损失率达到 9.77%。
关键词:葡萄病毒病;卷积神经网络;模型;图像识别
DOI:10.19850/j.cnki.2096-4706.2021.10.007
中图分类号:TP391.4 文献标识码:A 文章编号:2096-4706(2021)10-0030-04
Image Identification of Grapevine Virus Disease Based on Convolution Neural Network
WU Xueqi
(Ji Xianlin Honors School of Liaocheng University,Liaocheng 252000,China)
Abstract:Grapes have a long history in the world. With the development of science and technology and the movement of population,the planting range of grape has been continuously expanded,ranking first in the world. However,this also brings many problems,such as grapevine virus disease. Early detection of diseases can reverse the downward trend of grape yield. Aiming at the problem of low judgment accuracy of human eyes,this paper proposes the image identification model of grapevine virus disease based on CNN,including an input layer,four convolution layers,four pooling layers,two fully connected layers and one output layer. For the data selected in this paper,the accuracy of this model reaches 97.25%,and the loss rate is 9.77%.
Keywords:grapevine virus disease;convolutional neural network;model;image identification
参考文献:
[1] 李晨,范旭东,张尊平,等 . 基于 RT-PCR 检测对 6 个 香味葡萄品种感染病毒状况的分析 [J]. 中国果树,2021(2):66- 69.
[2] 龙满生,欧阳春娟,刘欢,等 . 基于卷积神经网络与迁移 学习的油茶病害图像识别 [J]. 农业工程学报,2018,34(18): 194-201.
[3] 杜海顺,蒋曼曼,王娟,等 . 一种用于农作物叶部病害图 像识别的双权重协同表示分类方法 [J]. 计算机科学,2017,44 (10):302-306+311.
[4] 谢觉,唐俊 . 基于视频流和位置流混合的建筑施工人员行 为识别研究 [J]. 电子世界,2019(16):49-50.
[5] 吴进,闵育,李聪,等 . 一种基于 3D-CNN 的微表情识 别算法 [J]. 电讯技术,2019,59(10):1115-1120.
作者简介:吴雪琦(2001—),女,汉族,山东阳谷人,本科 在读,研究方向:模式识别。