摘要
为准确有效识别出农作物病虫害类别及位置,构建一款农作物病虫害图像识别App系统,为广大农户、研究人员及管理者提供智能信息服务。该系统基于Android平台开发,在所收集的大量病虫害数据集上,开展了Darknet、YOLO等深度网络模型训练和测试,并使用批量正则化、维度聚类和课程设计学习等技术优化模型,实现了181种作物病虫害图像的在线识别检测,为复杂环节下农作物病害及虫害在线识别、监管防控、综合治理等提供技术依据。
In order to accurately and effectively identify the types and locations of crop diseases and insect pests,we construct a crop disease and insect image recognition App system,which provides intelligent information services for farmers,researchers and managers.We developed the system on Android platform.Darknet,YOLO and other deep network model training and testing were carried out on a large number of collected pest and disease data set.The model was optimized by using batch regularization,dimensional clustering,and course design learning.It realized the online recognition and detection of 181 crop diseases and insect pests images,and provides technical basis for the online recognition,supervision and control,and comprehensive management of crop diseases and pests under complex links.
作者
陶治
孔建磊
金学波
白玉廷
苏婷立
Tao Zhi;Kong Jianlei;Jin Xuebo;Bai Yuting;Su Tingli(School of Computer and Information Engineering,Beijing Technology and Business University,Beijing 100048,China;Beijing Key Laboratory of Big Data Technology for Food Safety,Beijing Technology and Business University,Beijing 100048,China)
出处
《计算机应用与软件》
北大核心
2022年第3期341-345,共5页
Computer Applications and Software
基金
国家重点研发计划项目(2017YFC1600605)
国家自然科学基金项目(61673002,61903009)
北京市教育委员会科技计划一般项目(KM201910011010)。
