摘要
目的:随着农业现代化的发展,精准喷施农药等数字农业,需要精确的农作物病害信息以及位置信息。方法:为满足现代化农业的精准喷药需求,首先提出一种农作物病害叶片语义分割模型,该模型在提供病害类别的同时提供像素级别的位置;然后基于该模型,利用树莓派4B微型电脑开发精准喷药小车;利用语意分割模型对农作物病害进行识别。结果:在PlantVillage数据集上,本研究设计模型精确度和平均交并比分别高度达93.1%和86.6%。结论:基于该模型,以树莓派4B为工控办开发了精准喷药智能小车,为研究应用落地提供基础。
Objective:With the development of agricultural modernization,precise pesticide spraying and other digital agriculture needs accurate crop disease information and location information.Methods:To meet the demand of precision spraying in modern agriculture,this paper proposed a semantic segmentation model of crop disease leaves was proposed,which provides the location of pixel level while providing disease categories.Then,based on this model,the precision spraying car was developed by using Raspberry Pi 4B microcomputer.Results:Finally the semantic segmentation model was used to identify crop diseases and on the PlantVillage dataset,the accuracy and average intersection ratio of the design model were 93.1%and 86.6%.Conclusion:Based on the model,a precision spraying intelligent car is developed with Raspberry Pi 4B as the industrial control office,which provides the basis for research and application.
作者
刘彬
杨选将
钟昌源
胡泽林
郭盼盼
李华龙
刘先旺
李淼
LIU Bin;YANG Xuanjiang;ZHONG Changyuan;HU Zelin;GUO Panpan;LI Hualong;LIU Xianwang;LI Miao(Anhui Jianzhu University,Hefei 230601,China;Institute of Intelligent Machines,Hefei Institutes of Physical Science,Chinese Academy of Sciences,Hefei 230031,China;University of Science and Technology of China,Hefei 230026,China)
出处
《安徽科技学院学报》
2021年第3期7-14,共8页
Journal of Anhui Science and Technology University
基金
广东省科技计划项目(2019B020216001)
安徽省重点研究与开发计划项目(202004e11020011)
国家自然基金项目(31902205)
安徽省重点研究和开发计划项目(201904a06020041)。
关键词
病害识别
神经网络
语义分割
精准喷药
智能小车
Disease recognition
Neural network
Semantic segmentation
Precision spraying
Intelligent car
作者简介
刘彬(1988-),男,安徽当涂人,硕士,助理工程师,主要从事农业工程研究;通信作者:李淼,研究员,博导,E-mail:mli@iim.ac.cn。
