摘要
针对杂草的精确喷洒问题提出一种基于卷积神经网络(Convolution Neural Network,CNN)的棉花植株和杂草的检测识别方法。首先采集不同环境下棉田中棉花植株和不同种类的杂草图像作为网络模型的数据集,对数据集进行数据增强来增加数据集的数量,将其分为训练集与测试集;然后构建CNN模型,在模型中添加Dropout层,以防止网络出现过拟合,将训练集数据输入网络模型,使模型学习棉花植株和杂草的特征信息;最后将测试集数据输入CNN模型,测试CNN模型对棉花植株和杂草的识别能力。研究结果表明CNN对于棉花植株和杂草的分类结果精度超过了99.95%,识别时间为197.2s,证明CNN可以快速高效的识别棉田中棉花植株和杂草,为农业智能精确除草装备的研发提供研究基础。
Aiming at the problem of accurate weed spraying,a cotton plant and weed detection and recognition method based on Convolution Neural Network(CNN)was proposed.Firstly,images of cotton plants and weeds of different species in cotton fields under different environments were collected as data sets of network model,and data sets were enhanced to increase the number of data sets,which were divided into training sets and test sets.Then,a CNN model was constructed,and a Dropout layer was added to the model to prevent overfitting.The training set data was input into the network model to make the model learn the characteristic information of cotton plants and weeds.Finally,the data of the test set were input into the CNN model to test the identification ability of the model to cotton plants and weeds.The results show that the classification accuracy of CNN for cotton plants and weeds exceeds 99.95%,and the recognition time is 197.2s,which proves that CNN can quickly and efficiently identify cotton plants and weeds in cotton fields,and provides a research basis for the research and development of agricultural intelligent and accurate weeding equipment.
作者
姚思雨
王磊
张宏文
YAO Siyu;WANG Lei;ZHANG Hongwen(College of Mechanical and Electrical Engineering,Shihezi University,Shihezi,Xinjiang 832003,China)
出处
《石河子大学学报（自然科学版）》
CAS
北大核心
2023年第1期21-26,共6页
Journal of Shihezi University(Natural Science)
基金
兵团重点领域创新团队(2019CB006)。
关键词
棉花植株
杂草识别
深度学习
卷积神经网络
cotton plant
weed identification
deep learning
convolutional neural networks
