当前位置>主页 > 期刊在线 > 计算机技术 >

计算机技术21年23期

YOLOv3 在安全帽佩戴检测中的应用
唐勇 ¹,巫思敏 ²
(1. 广东工业大学,广东 广州 510006;2. 北方民族大学 计算机科学与工程学院,宁夏 银川 750021)

摘  要:在工地现场的安全管理中,施工人员的安全帽佩戴与否关系到其生命安全。传统的人工方式检测施工人员安全帽佩戴与否的效率较低。文章提出一种基于 YOLOv3 目标检测算法,可以实时地自动检测施工人员是否佩戴安全帽:首先收集安全帽数据集和标注处理,把数据集分为训练集、验证集和测试集;然后使用 YOLOv3 对训练集和验证集进行训练,并对测试集进行测试,最终结果显示检测准确率达到 0.78、召回率 0.914、mAP 达到 0.91、F1 指数 0.842、FPS 为 161,可满足安全帽实时检测的要求。


关键词:YOLOv3;目标检测;安全帽



DOI:10.19850/j.cnki.2096-4706.2021.23.023


中图分类号:TP391                                     文献标识码:A                                  文章编号:2096-4706(2021)23-0088-05


Application of YOLOv3 in Safety Helmet Wearing Detection

TANG Yong1 , WU Simin2

(1.Guangdong University of Technology, Guangzhou 510006, China; 2.School of Computer Science and Engineering, North Minzu University, Yinchuan 750021, China)

Abstract: Whether constructors wear safety helmets or not is related to their life safety in the safety management of the construction site. It is inefficient to detect whether the safety helmet of construction personnel is worn by traditional manual inspection. This paper proposes a target detection algorithm based on YOLOv3 to detect automatically whether the construction personnel wear safety helmets in real time. Firstly, collect the data set of safety helmets and label processing, divide the data set into training set, verification set and test set; then use YOLOv3 to train the training set and verification set, and test the test set. The final results show that the accuracy rate of detection reaches 0.78, the recall rate is 0.914, mAP reaches 0.91, F1 index is 0.842, and FPS is 161, which can meet the real-time detection requirements of safety helmet.

Keywords: YOLOv3; target detection; safety helmet


参考文献:

[1] 唐凯,陈陆,张洲境,等.我国建筑施工图 4 损失函数走势图行业生产安全事故统计分析及对策 [J].建筑安全, 2020,35(9):40-43.

[2] 谢林江,季桂树,彭清,等.改进的卷积神经网络在行人检测中的应用 [J].计算机科学与探索,2018,12(5):708-718.

[3] 彭清,季桂树,谢林江,等.卷积神经网络在车辆识别中的应用 [J].计算机科学与探索,2018,12(2):282-291.

[4] 冯国臣,陈艳艳,陈宁,等.基于机器视觉的安全帽自动识别技术研究 [J].机械设计与制造工程,2015,44(10):39- 42.

[5] REDMON J,DIVVALA S,GIRSHICK R,et al. You Only Look Once:Unified,Real-Time Object Detection [C]//2016 IEEE Conference on Computer Vision and Pattern Recognition(CVPR).Las Vegas:IEEE,2016:779-788.

[6] REDMON J,FARHADI A. YOLO9000:Better,Faster,Stronger [C]//2017 IEEE Conference on Computer Vision and Pattern Recognition(CVPR).Honolulu:IEEE,2017:6517-6525.

[7] REDMON J,FARHADI A. YOLOv3:An Incremental Improvement [J/OL].arXiv:1804.02767 [cs.CV].[2021-11-02].https:// arxiv.org/abs/1804.02767.

[8] NEUBECK A,GOOL L V. Efficient Non-Maximum Suppression [C]//8th International Conference on PatternRecognition (ICPR06).Hong Kong:IEEE,2006:850-855.


作者简介:唐勇(1993—)男,汉族,广东河源人,硕士研究生在读,主要研究方向:深度学习与目标检测;通信作者:巫思敏 (1993—)女,汉族,广东信宜人,硕士研究生,高级工程师,主要研究方向:数据挖掘与知识发现。