基于YOLOv7模型的无核白葡萄分级系统研究
Study on classification system for seedless white grape based on YOLOv7 model
李园园 1周文静 1崔振宇 1卜露1
作者信息
- 1. 新疆科技学院信息科学与工程学院,新疆库尔勒 841000
- 折叠
摘要
无核白葡萄品质的参差不齐导致了其市场竞争力差,人工分级效率低且存在主观差异,本文提出了一种基于YOLOv7模型的无核白葡萄分级系统.将无核白葡萄果实图像作为模型的输入,对YOLOv7模型进行训练和调整所得模型的平均精度高达96.86%,分级平均速度达1.9张/s.将本文的分级模型与传统的人工分级方式进行对比,验证了该模型在对不同品质无核白葡萄果实分级中的优势.试验结果表明,基于YOLOv7模型的无核白葡萄分级系统可以实时、准确地对不同级别无核白葡萄果实进行分级,并且在识别精度、速度等方面均优于传统的人工分级方式,该系统可为无核白葡萄果实分级研究提供参考.
Abstract
The uneven quality of seedless white grapes leads to poor market competitiveness,while manual grading has low effi-ciency and subjective differences.Therefore,the objective of this paper is to propose a seedless white grape classification system based on the YOLOv7 model.The image of the seedless white grape fruit was used as input to train and adjust the YOLOv7 model,resulting in an average accuracy of 96.86%and an average grading speed of 1.9 images per second.The grading model proposed in this article was compared with traditional manual grading methods to verify its advantages in grading seedless white grape fruits of different qualities.The experimental results show that the YOLOv7 model can distinguish different levels of seedless white grapes in real time and accurately,and is superior to traditional manual grading methods in terms of grading accuracy and speed.This classification system provides a valuable reference for research of fruit grading in seedless white grapes.
关键词
YOLOv7/无核白葡萄/分级检测Key words
YOLOv7/Seedless white grapes/Classification detection引用本文复制引用
基金项目
国家级大学生创新创业训练计划(202213561004)
自治州科学技术研究计划(202201)
出版年
2024
NETL
NSTL
万方数据