摘要
玉米叶片是玉米进行光合作用的主要器官,其生理生态的快速检测对玉米的质量和产量起着至关重要的作用。因此,文章提出一种基于YOLOv5模型的玉米叶片识别方法。文章利用Kinect 2.0相机水平获取了玉米植株生长时期的图像数据;利用YOLOv5模型,实现了玉米叶片目标检测,并将其与YOLOv3和YOLOv7进行对比。结果表明:YOLOv5模型的精确度、召回率、F1分数分别达到了94.7%、95.2%和93.4%,其平均精度均值与YOLOv3和YOLOv7相比提高了3.9%、1.4%。该成果为玉米叶片实时检测系统的研发注入了强大动力,同时也在人工智能育种和估产领域展现了不可或缺的技术支撑作用。
Corn leaf is the main organ of corn for photosynthesis,and the rapid detection of its physiological ecology plays a crucial role in the quality and yield of corn.Therefore,this paper proposes a corn leaf recognition method based on the YOLOv5 model.The image data of corn plant growth period is acquired by using Kinect 2.0 camera level,and by using YOLOv5 model,corn leaf target detection is realized and compared with YOLOv3 and YOLOv7.The result shows that the precision,recall,and F1-score of the YOLOv5 model reached 94.7%,95.2%,and 93.4%,respectively,and compared with YOLOv3 and YOLOv7,its mean average precision value is improved by 3.9%and 1.4%.The results injects strong impetus to the research and development of real-time detection system for corn leaf,and also demonstrates its indispensable technical support in the field of artificial intelligence breeding and estimation yield.
作者
罗苏
关海鸥
LUO Su;GUAN Hai’ou(College of Information and Electrical Engineering,Heilongjiang Bayi Agricultural University,Daqing 163319,China)
出处
《无线互联科技》
2024年第16期23-26,44,共5页
Wireless Internet Science and Technology
基金
国家重点研发计划,项目名称:优质专用、耐密高产、宜粒收玉米品种筛选标准体系构建,项目编号:2023YFD2301701。
