摘要
植物表型是植物受自身基因表达、环境等影响后的外在表现,其决定了农作物产量、品质和抗逆性等。大多数植物表型信息可以通过数字图像处理的方法来获取和分析,且随着深度学习在数字图像处理领域的进展,基于深度学习的图像处理在技术表现上远胜于传统方法。深度学习在网络中有更多的隐藏层,具有更大的鉴别能力和预测能力。文中使用深度学习中的卷积神经网络来自动识别数据集中的定量特征,验证了该方法在小麦植物表型识别、分类、特征识别和定位中的高精度应用。
Plant phenotype is the external performance of plants affected by their own gene expression and environment,which determines crop yield,quality and stress resistance.Most plant phenotype information can be obtained and analyzed by digital image processing methods,and with the progress of deep learning in the field of digital image processing,deep learning-based image processing is far better than traditional methods in technical performance.Deep learning has more hidden layers in the network and has greater discriminative and predictive capabilities.In this paper,convolutional neural networks in depth learning are used to automatically identify quantitative features in datasets,and the application of this method in wheat plant phenotype recognition,classification,feature recognition and localization with high precision is veri-fied.
作者
张举
ZHANG Ju(School of Communication Engineering,Taishan College of Science and Technology,Tai'an,Shandong 271038,China)
出处
《移动信息》
2023年第11期132-135,共4页
MOBILE INFORMATION
关键词
植物表型
深度学习
图像分析
卷积神经网络
Plant phenotype
Deep learning
Image analysis
Convolutional neural networks