摘要
育种因周期长、难度大、失败率高使得选育优良品种进程较为缓慢,基于卷积神经网络的图像识别技术因占用资源少、准确率高的优点而得到了大量应用。本文阐释了基于卷积神经网络的图像识别技术的基本原理,并以识别统计水稻稻粒为例进行了实际运用展示,论证了基于卷积神经网络的图像识别技术在育种工作中的可行性。随着技术进一步成熟,未来大量性状分辨记录、数量统计等工作将由机器完成,减轻育种工作的工作量。
Due to the long cycle,difficulty and high failure rate of breeding,the process of selecting superior varieties is relatively slow.With the rapid development of image recognition technology based on convolutional neural networks,which has the advantages of low resource consumption and high accuracy.This paper explains the basic principle of image recognition technology based on convolutional neural network and takes the recognition and statistics of rice grains as an example to demonstrate the practical application,and demonstrates the feasibility of image recognition technology based on convolutional neural network in breeding work.In the future,with the further maturity of technology,a large amount of repetitive work such as character discrimination records and quantitative statistics will be completed by machines,which will reduce a lot of workload for breeding work.
作者
刘洋
张钊
夏旭
韩学坤
Liu Yang;Zhang Zhao;Xia Xu;Han Xuekun(HuaiHua Agricultural Sciences Academy of Dryland Crops Research Institute,Huaihua 418000,China)
出处
《现代农业装备》
2023年第3期57-60,共4页
Modern Agricultural Equipment
关键词
育种
卷积神经网络
图像识别技术
水稻稻粒统计
breeding
convolutional neural networks
image recognition technology
statistics of rice grains