面向复杂施工环境的人员防护装备穿戴识别方法对比研究

荣耀; 于平林; 李永威; 王培; 丁海滨

doi:10.13637/j.issn.1009-6094.2025.081410.13637/j.issn.1009-6094.2025.0814

Chinese

您当前的位置：

首页 >

文章列表页 >

面向复杂施工环境的人员防护装备穿戴识别方法对比研究

更新时间：2026-04-17

- 面向复杂施工环境的人员防护装备穿戴识别方法对比研究
- Vol. 26, Issue 4, Pages: 1328-1337(2026)
- 作者机构：
  
  1. 江西省交投养护科技集团有限公司
  2. 华东交通大学山区土木工程安全与韧性全国重点实验室
- 作者简介：
- 基金信息：
- DOI：10.13637/j.issn.1009-6094.2025.081410.13637/j.issn.1009-6094.2025.0814
  CLC： TP391.41;TU714
- Published：2026
- 稿件说明：
移动端阅览
[1]荣耀,于平林,李永威,等.面向复杂施工环境的人员防护装备穿戴识别方法对比研究[J].安全与环境学报,2026,26(04):1328-1337.
[1]荣耀,于平林,李永威,等.面向复杂施工环境的人员防护装备穿戴识别方法对比研究[J].安全与环境学报,2026,26(04):1328-1337. DOI： 10.13637/j.issn.1009-6094.2025.081410.13637/j.issn.1009-6094.2025.0814.

DOI：10.13637/j.issn.1009-6094.2025.0814.

摘要

施工现场环境复杂，传统巡检手段准确率低。为解决这一问题，通过数据增强提升模型泛化能力，并对比YOLO系列算法，确定适用于现场的装备识别方法，拓展至视频检测，实现高效识别。结果表明：(1)通过光照增强、噪声模拟及数据清洗，将原始图像从1 395张扩展至1 795张，有效提升了模型适应性；(2)对比了不同算法在安全帽与反光衣穿戴识别的准确度，发现YOLOv11n在P、R和mAP50三项指标中表现最优，分别达到0.869、0.806、0.830

相较于基础模型YOLOv8

改进后的YOLOv8n在P、R、mAP50和mAP50-95上分别提升了6.2百分点、8.7百分点、8百分点和11百分点，相较于基础模型YOLOv11

YOLOv11n则分别提升了6.8百分点、12百分点、10.7百分点和12.3百分点，表明两者在检测精度与综合性能方面均有显著改进；(3)通过对比分析表明YOLOv8n在昏暗环境与视频处理中的综合表现最佳。研究成果为复杂施工现场智能安全管理奠定了基础。

Abstract

This study presents a comparative method for detecting the wearing status of Personal Protective Equipment(PPE)

such as safety helmets and reflective vests

in complex construction environments. The approach utilizes various versions of the YOLO-series object detection algorithms and incorporates data augmentation techniques to enhance model performance. The original dataset comprised 1 395 images

which were augmented by adjusting brightness

contrast

saturation

and exposure

as well as by introducing artificial noise. After a rigorous data cleaning process

the dataset was expanded to 1 795 images

ensuring both high quality and diversity. Seven YOLO models—YOLOv3

YOLOv5

YOLOv8

YOLOv8n

YOLOv8l

YOLOv11

and YOLOv11n—were trained within a unified framework. The input image resolution was set to 640 pixels×640 pixels

with a batch size of 4. To prevent overfitting

an early stopping mechanism(patience=100) was implemented. Training took place on a Windows 10 system equipped with an NVIDIA GeForce MX450 GPU

utilizing the PyTorch framework. Model performance was evaluated using precision(P)

recall(R)

mean Average Precision at IoU 0.50(mAP50)

and mean Average Precision at IoU 0.50 to 0.95(mAP50-95). The results indicated that YOLOv11n achieved the best performance with a precision of 0.869

recall of 0.806

and mAP50 of 0.830. However

YOLOv8n slightly outperformed in mAP50-95

achieving a score of 0.638

which suggests stronger adaptability to varying IoU thresholds. In low-light environments

YOLOv8n demonstrated the highest accuracy in image detection

recording no false positives or missed detections. For video-based detection

frames were extracted every 0.5 seconds

and YOLOv8n once again exhibited the most stable performance

closely followed by YOLOv11n

which missed only two instances. In contrast

earlier models such as YOLOv3 and YOLOv5 demonstrated lower stability and accuracy

particularly when detecting small objects or in cluttered scenes. In summary

the enhanced YOLOv8n and YOLOv11n models showcased excellent robustness and detection accuracy in complex construction environments

supporting their deployment in real-time intelligent safety monitoring systems on construction sites.

关键词

Keywords

references

VIOLA P,JONES M.Rapid object detection using a boosted cascade of simple features[C]//Proceedings of the 2001 Institute of Electrical and Electronics Engineers Computer Society Conference on Computer Vision and Pattern Recognition.Piscataway,NJ,USA:Institute of Electrical and Electronics Engineers,2001:511-518.

LOWE D G.Distinctive image features from scale-invariant keypoints[J].International Journal of Computer Vision,2004,60(2):91-110.

DALAL N,TRIGGS B.Histograms of oriented gradients for human detection[C]//Proceedings of the 2005 Institute of Electrical and Electronics Engineers Computer Society Conference on Computer Vision and Pattern Recognition.Piscataway,NJ,USA:Institute of Electrical and Electronics Engineers,2005:886-893.

KRIZHEVSKY A,SUTSKEVER I,HINTON G E.ImageNet classification with deep convolutional neural networks[C]//Advances in Neural Information Processing Systems 25 (NIPS 2012).Red Hook:Curran Associates Inc.,2012:1097-1105.

REDMON J,DIVVALA S,GIRSHICK R,et al.You Only Look Once:unified,real-time object detection[C]//Proceedings of the 2016 Institute of Electrical and Electronics Engineers Conference on Computer Vision and Pattern Recognition.Piscataway,NJ,USA:Institute of Electrical and Electronics Engineers,2016:779-788.

REDMON J,FARHADI A.YOLO9000:better,faster,stronger[C]//Proceedings of the 2017 Institute of Electrical and Electronics Engineers Conference on Computer Vision and Pattern Recognition.Piscataway,NJ,USA:Institute of Electrical and Electronics Engineers,2017:7263-7271.

王秋余.基于视频流的施工现场工人安全帽佩戴识别研究[D].武汉：华中科技大学，2018.WANG Q Y.Research on safety helmet wearing recognition for construction workers based on video stream[D].Wuhan:Huazhong University of Science and Technology,2018.

DELHI V S K,SANKARLAL R,THOMAS A.Detection of personal protective equipment (PPE) compliance on construction site using computer vision based deep learning techniques[J].Frontiers in Built Environment,2020,6:136.

邱云飞，腰瑞琳，金海波，等.HD-YOLO:复杂场景下安全帽佩戴检测算法[J].安全与环境学报，2025,25(1):165-174.QIU Y F,YAO R L,JIN H B,et al.HD-YOLO:safety helmet detection algorithm in complex scenes[J].Journal of Safety and Environment,2025,25(1):165-174.

谢国波，肖峰，林志毅，等.复杂作业场景下的反光衣和安全帽检测方法[J].安全与环境学报，2024,24(9):3513-3521.XIE G B,XIAO F,LIN Z Y,et al.Method for detecting reflective vests and safety helmets in complex operational environments[J].Journal of Safety and Environment,2024,24(9):3513-3521.

Views

201

下载量

CSCD

Alert me when the article has been cited

提交

Tools

Publicity Resources

LRM-YOLO:一种面向工业现场的轻量化安全帽佩戴检测方法

复杂场景下融合多尺度特征与注意力机制的安全帽检测

基于YOLOv10n-knot的电力安全带挂点检测算法

改进型金字塔解析网络在矿区浮石识别中的应用研究

沿空留巷采空区注氮防灭火参数优化

Related Author

张新君

王贺桐

张永库

邱云飞

齐焓玮

金海波

郑艳梅

张欣

Related Institution

辽宁工程技术大学电子与信息工程学院

辽宁工程技术大学辽宁省无线射频大数据智能应用重点实验室

辽宁工程技术大学软件学院

河南理工大学计算机科学与技术学院

西安建筑科技大学资源工程学院

AI问答

Postal code：100079
Tel：（010）53879206 Email：tmw@bjxintong.com.cn
Technical support is provided by Beijing Founder electronics co., LTD 京ICP备09082226号-64 京公网安备11010602201714号
It is recommended to read the content of this site in Chrome&IE9+. Please switch to extreme mode in browser 360.
Cookies We use cookies to help provide and enhance our service and tailor content. By continuing, you agree to the use of cookies.

⁰