摘要
花椒簇的准确识别是花椒采摘机器人研制的关键,基于YOLOv8n模型对花椒簇图像进行了识别。采集构建了花椒簇图像数据集并进行了数据增强,利用YOLOv8n模型进行了训练,模型的平均精度均值mAP@0.5和mAP@0.5~0.95分别为0.881和0.622,模型参数量和浮点运算量FLOPs分别为3.011 M和8.7 G,较好地实现了模型准确性与复杂度的平衡,为花椒采摘机器人的研制提供支持。
Accurate identification of Zanthoxylum bungeanum clusters is critical to the development of Zanthoxylum bungeanumpicking robots.In this paper,the YOLOv8n model was used to identify images of Zanthoxylum bungeanum clusters.The dataset of Zanthoxylum bungeanum clusters images was collected and constructed,followed by data augmentation.The YOLOv8n model was trained using this dataset,achieving mean average precision mAP@0.5 and mAP@0.5~0.95 of 0.881 and 0.622,respectively.The model has 3.011 M parameters and 8.7 G FLOPs,which provided a well-balanced accuracy and complexity.This result could provide support for the development of anthoxylum bungeanum-picking robots.
作者
王硕
张志勇
张燕青
乔宇
武同辉
Wang Shuo;Zhang Zhiyong;Zhang Yanqing;Qiao Yu;Wu Tonghui(College of Agricultural Engineering,Shanxi Agricultural University,Taigu 030801,Shanxi,China)
出处
《农业技术与装备》
2025年第12期3-5,共3页
Agricultural Technology & Equipment
基金
山西省专利转化专项计划项目(202301012)。
关键词
花椒簇
YOLOv8n模型
识别
Zanthoxylum bungeanum clusters
YOLOv8n model
identification
