• 首页
  • 期刊简介
  • 编委会
  • 投稿指南
  • 收录情况
  • 杂志订阅
  • 联系我们
引用本文:杨树元,陶 卓,刘天萌,李 昌,赵子钧.基于YOLOv8的实时教室人数统计系统[J].软件工程,2024,27(12):1-4.【点击复制】
【打印本页】   【下载PDF全文】   【查看/发表评论】  【下载PDF阅读器】  
←前一篇|后一篇→ 过刊浏览
分享到: 微信 更多
基于YOLOv8的实时教室人数统计系统
杨树元1, 陶 卓2, 刘天萌2, 李 昌2, 赵子钧2
(1.河北师范大学数学科学学院, 河北 石家庄 050024;
2.河北师范大学软件学院, 河北 石家庄 050024)
368392@qq.com; taozhuo@hebtu.edu.cn; liutianmeng@onest.net; 1623334379@qq.com; 2094555264@qq.com
摘 要: 为了实现教室使用率和学生出勤情况的高效管理,推动教务管理的自动化进程,设计并实现了一种基于YOLOv8的端到端实时教室人数统计系统。此系统仅需利用教室内的普通摄像头,即可实现对教室人数的实时查询与可视化展示。为提升模型对特定教室环境的适应性,加入教室摄像头自制数据集ClassCount对YOLOv8模型进行微调。实验结果表明,加入ClassCount数据集进行训练后,模型在混合数据集上的表现显著提升,其中mAP@50提高了约7.7%,mAP@50~95提高了约6.8%,验证了模型在复杂教室环境中进行实时人数统计时的适应性和鲁棒性。
关键词: YOLOv8;人数统计;深度学习;目标检测
中图分类号: TP311    文献标识码: A
基金项目: 河北省教育厅2021—2022年度河北省高等教育教学改革研究与实践项目:基于机器学习的智能教辅系统实训项目开发与应用研究(2021GJJG123)
Real-Time Classroom Occupancy Counting System Based on YOLOv8
YANG Shuyuan1, TAO Zhuo2, LIU Tianmeng2, LI Chang2, ZHAO Zijun2
(1.School of Mathematical Sciences, Hebei Normal University, Shijiazhuang 050024, China;
2.School of So f tware, Hebei Normal University, Shijiazhuang 050024, China)
368392@qq.com; taozhuo@hebtu.edu.cn; liutianmeng@onest.net; 1623334379@qq.com; 2094555264@qq.com
Abstract: To achieve efficient management of classroom utilization and student attendance, and to promote the automation of educational management processes, this paper proposes to design and implement an end-to-end real-time classroom occupancy counting system based on YOLOv8. This system can perform real-time inquiries and visual display of classroom occupancy using only ordinary cameras installed within the classroom. To enhance the model's adaptability to the specific classroom environment, a custom dataset called ClassCount is created to fine-tune the YOLOv8 model. Experimental results indicate that after training with the ClassCount dataset, the model's performance on a mixed dataset significantly improves, with mAP @ 50 increasing by approximately 7.7% and mAP @ 50~95 improving by around 6.8% . This verifies the model's adaptability and robustness for real-time occupancy counting in complex classroom environment.
Keywords: YOLOv8; occupancy counting; deep learning; object detection


版权所有:软件工程杂志社
地址:辽宁省沈阳市浑南区新秀街2号 邮政编码:110179
电话:0411-84767887 传真:0411-84835089 Email:semagazine@neusoft.edu.cn
备案号:辽ICP备17007376号-1
技术支持:北京勤云科技发展有限公司

用微信扫一扫

用微信扫一扫