铁路肩负着运送大量货物和旅客的重任,危险性较高,铁路工作人员若不正确佩戴安全装备,则发生事故的概率会显著升高,因此提升铁路安全装备检测精度十分必要。为了预防铁路作业中因铁路安全装备缺失导致的安全事故,提升工人安全帽和安全背心佩戴情况的智能检测能力,以YOLOv8为基础模型提出一个RST-YOLOv8模型,在主干网络替换RepViT模块,提高安全帽和反光背心的检测精度;在颈部网络使用SSFF模块和TFE模块,提升对小目标的检测效果。实验结果表明,改进后模型相较原YOLOv8模型,mAP50 和mAP50-95分别提高1.9%和1.7%。
Abstract
The railway is responsible for the transport of large quantities of goods and passengers, and the danger is high, if railway staff do not wear safety equipment correctly, the probability of accidents will be significantly increased, so it is necessary to improve the detection accuracy of railway safety equipment. In order to prevent safety accidents caused by the lack of railway safety equipment in railway operations and improve the intelligent detection ability of workers' safety helmets and safety vests, this paper proposes an RST-YOLOv8 model based on YOLOv8 model, which replaces the RepViT module in the backbone network to improve the detection accuracy of helmet and vests. SSFF module and TFE module are used in the neck network to improve the detection effect of small targets. Experimental results show that compared with the original YOLOv8 model, the improved model increases mAP50 and mAP50-95 by 1.9 % and 1.7 %, respectively.
关键词
安全装备检测 /
卷积神经网络 /
YOLOv8 /
铁路安全
Key words
safety equipment testing /
convolutional neural networks /
YOLOv8 /
railway safety
{{custom_sec.title}}
{{custom_sec.title}}
{{custom_sec.content}}
参考文献
[1] 施辉,陈先桥,杨英.改进YOLO v3的安全帽佩戴检测方法[J].计算机工程与应用,2019,55(11):213-220.
[2] 吴冬梅,尹以鹏,宋婉莹,等.改进YOLO v4算法的电动车驾驶员头盔佩戴检测[J].计算机仿真,2023,40(3):508-513.
[3] 张锦,屈佩琪,孙程,等.基于改进YOLOv5的安全帽佩戴检测算法[J].计算机应用,2022,42(4):1292-1300.
[4] Dihua W.Using Channel Pruning-Based Yolo V4 Deep Learning Algorithm For The Real-Time And Accurate Detection Of Apple Flowers In Natural Environments[J].Computers and Electronics in Agriculture,2020 (178):105742.
[5] Dandan W,Dongjian H.Channel Pruned Yolo V5s-Based Deep Learning Approach For Rapid And Accurate Apple Fruitlet Detection Before Fruit Thinning[J].Biosystems engineering,2021 (210):271-281.
[6] Rui S.An attribution-based pruning method for real-time mango detection with YOLO network[J].Computers and Electronics in Agriculture,2020(169):105214.
[7] Aboah A,Wang B,Bagci U,et al.Real-time multi-class helmet violation detection using few-shot data sampling technique and YOLOv8[C]//2023 IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops (CVPRW).Vancouver,2023:5350-5358.
[8] WANG A,CHEN H,LIN Z,et al.Repvit:revisiting mobileCNN from vit perspective[C]//Proceedings of the IEEE/CVF conference on computer vision and pattern recognition.Nashville:IEEE,2024:15909-15920.
[9] 王盘龙,严云洋,耿嘉雯,等.基于改进YOLOv9的火焰检测算法[J].扬州大学学报(自然科学版),2025,28(1):64-72.
[10] 董善,陈清江.基于多尺度特征融合的复杂交通场景目标检测算法[J/OL].无线电工程,1-12[2025-04-09].
[11] 何梦珂,牟莉,李文卓.基于深度学习的小目标工件检测算法[J].计算机技术与发展,2025,35(2):41-47.
[12] KANG M,TING C M,TING F F,et al.ASF-YOLO:a novel YOLO model with attentional scale sequence fusion for cell instance segmentation[J].Image and Vision Computing,2024,147:105057.
[13] 胡卫昊,唐超礼.轻量级YOLOv8在复杂天气下的绝缘子缺陷检测[J/OL].重庆工商大学学报(自然科学版),1-10[2025-04-09].
[14] Nie H,Pang H,Ma M,et al.A Lightweight Remote Sensing Small Target Image Detection Algorithm Based on Improved YOLOv8[J].Sensors,2024,24(9):2952-.
[15] He Y,Sahma A,He X,et al.FireNet:A Lightweight and Efficient Multi-Scenario Fire Object Detector[J].Remote Sensing,2024,16(21):4112-4112.
