|
|
|
| Improved railway worker safety wear detection model based on YOLOv11n |
Xiaochun WU( ),Zining LI |
| School of Automation and Electrical Engineering, Lanzhou Jiaotong University, Lanzhou 730070, China |
|
|
|
Abstract A VMS-YOLOv11n model for railroad safety wear detection was proposed aiming at the problems of more complex railroad operation scenario, smaller safety wear vision targets and limited resource of edge computing devices. The VABlock feature extraction module was designed to replace the C3k2 module in the backbone network as well as the neck network to reduce computational redundancy while retaining key features. Then the MSF multi-scale feature fusion structure was designed to improve the small target detection accuracy in complex railroad scene by fusing the multilevel feature weighting strategy with the bidirectional feature pyramid network. The spatial pyramid feature fusion structure (SE2SPPF) was constructed to enhance the small target feature learning capability while improving the model detection capability. The experimental results showed that the improved model had precision rate of 98%, recall rate of 92.5%, and average precision mean of 97.5%, which were 2.9%, 2.5% and 2.7% higher, respectively compared with the baseline YOLOv11n. The number of parameters is 2.45×106, which is 5% lower than the baseline. The improved model works well in railroad safety wear detection, effectively balancing detection accuracy and speed.
|
|
Received: 26 September 2025
Published: 20 July 2026
|
|
|
| Fund: 国家自然科学基金资助项目(61661027)；中央引导地方科技发展资金资助项目(24ZYQA044)；甘肃省重点研发计划资助项目(22YF7GA141). |
基于YOLOv11n的改进铁路工人安全穿戴检测模型
针对铁路作业场景较复杂，安全穿戴远景目标较小及边缘计算设备资源有限等问题，提出VMS-YOLOv11n的铁路安全穿戴检测模型. 设计VABlock特征提取模块，替换主干网络及颈部网络中的C3k2模块，在保留关键特征的同时减少计算冗余. 设计MSF多尺度特征融合结构，通过融合多级特征加权策略与双向特征金字塔网络，提高铁路复杂场景下的小目标检测精度. 构建空间金字塔特征融合结构(SE2SPPF)，在增强小目标特征学习能力的同时提高模型检测能力. 实验结果表明，相较于基线YOLOv11n，改进模型的精确率为98%，召回率为92.5%，平均精度均值为97.5%，分别提高了2.9%、2.5%和2.7%. 参数量为2.45×106，与基线模型相比降低了5%. 改进后的模型在铁路安全穿戴检测中的效果良好，有效平衡了检测精度和速度.
关键词：
铁路安全穿戴检测,
YOLO,
复杂场景,
小目标,
特征提取
|
|
| [1] |
方文珊 《铁路轨道工程施工质量检测技术规程》主要技术标准研究[J]. 铁路工程技术与经济, 2025, 40 (1): 11- 14
FANG Wenshan Research on the major technical regulations in technical code for constructional quality detection of railway track[J]. Railway Engineering Technology and Economy, 2025, 40 (1): 11- 14
doi: 10.20262/j.cnki.issn.2097-6186.2025.01.03
|
|
|
| [2] |
周瑶, 周石 基于YOLOv5改进的铁路工人安全帽检测算法研究[J]. 计算机测量与控制, 2024, 32 (3): 71- 78
ZHOU Yao, ZHOU Shi Research on the detection algorithm of railway worker’s hard hat based on YOLOv5 improvement[J]. Computer Measurement and Control, 2024, 32 (3): 71- 78
doi: 10.16526/j.cnki.11-4762/tp.2024.03.011
|
|
|
| [3] |
REN S, HE K, GIRSHICK R, et al Faster R-CNN: towards real-time object detection with region proposal networks[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2017, 39 (6): 1137- 1149
doi: 10.1109/TPAMI.2016.2577031
|
|
|
| [4] |
HE K, GKIOXARI G, DOLLÁR P, et al. Mask R-CNN [C]//Proceedings of the IEEE International Conference on Computer Vision. Venice: IEEE, 2017: 2980–2988.
|
|
|
| [5] |
REDMON J, DIVVALA S, GIRSHICK R, et al. You only look once: unified, real-time object detection [C]//Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition. Piscataway: IEEE, 2016: 779-788.
|
|
|
| [6] |
LIU W, ANGUELOV D, ERHAN D, et al. SSD: single hot multiBox detector [C]//European Conference on Computer Vision. Cham: Springer, 2016: 21-37.
|
|
|
| [7] |
LING L, XU S, WEI L, et al Fsd-detr: casting surface defect detection based on improved RT-DETR[J]. Journal of Real-Time Image Processing, 2025, 22 (4): 135
doi: 10.1007/s11554-025-01714-x
|
|
|
| [8] |
冯勇, 杨思卓, 徐红艳 基于YOLO v8的轻量化安全帽佩戴检测算法[J]. 计算机应用, 2024, 44 (Suppl.2): 251- 256
FENG Yong, YANG Sizhuo, XU Hongyan Lightweight safety helmet wearing detection algorithm based on YOLO v8[J]. Journal of Computer Applications, 2024, 44 (Suppl.2): 251- 256
|
|
|
| [9] |
SHAN C, LIU H, YU Y Research on improved algorithm for helmet detection based on YOLOv5[J]. Scientific Reports, 2023, 13: 18056
doi: 10.1038/s41598-023-45383-x
|
|
|
| [10] |
徐壮, 钱育蓉, 颜丰 GCW-YOLOv8n: 轻量级安全帽佩戴检测算法[J]. 计算机工程与应用, 2025, 61 (3): 144- 154
XU Zhuang, QIAN Yurong, YAN Feng GCW-YOLOv8n: lightweight safety helmet wearing detection algorithm[J]. Computer Engineering and Applications, 2025, 61 (3): 144- 154
|
|
|
| [11] |
冯爽, 王万齐, 杨文, 等 基于改进RT-DETR的铁路施工场景下人员安全穿戴检测[J]. 铁道学报, 2025, 47 (2): 92- 101
FENG Shuang, WANG Wanqi, YANG Wen, et al Safety wear detection for personnel in railway construction scenarios based on improved RT-DETR[J]. Journal of the China Railway Society, 2025, 47 (2): 92- 101
doi: 10.3969/j.issn.1001-8360.2025.02.010
|
|
|
| [12] |
KHANAM R, HUSSAIN M. YOLOv11: an overview of the key architectural enhancements [EB/OL]. [2025-09-10]. https://arxiv.org/abs/2410.17725.
|
|
|
| [13] |
LI H, LI J, WEI H, et al. Slim-neck by GSConv: a lightweight-design for real-time detector architectures [EB/OL]. [2025-09-10]. https://arxiv.org/abs/2206.02424.
|
|
|
| [14] |
ZHANG X, SONG Y, SONG T, et al. AKConv: convolutional kernel with arbitrary sampled shapes and arbitrary number of parameters [EB/OL]. [2025-09-10]. https://arxiv.org/abs/2311.11587v1.
|
|
|
| [15] |
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
doi: 10.1016/j.imavis.2024.105057
|
|
|
| [16] |
CHEN J, KAO S H, HE H, et al. Run, don’t walk: chasing higher FLOPS for faster neural networks [C]//Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition. Vancouver: IEEE, 2023: 12021–12031.
|
|
|
| [17] |
程德强, 姬广凯, 张皓翔, 等 基于多粒度融合和跨尺度感知的跨模态行人重识别[J]. 通信学报, 2025, 46 (1): 108- 123
CHENG Deqiang, JI Guangkai, ZHANG Haoxiang, et al Cross-modality person re-identification based on multi-granularity fusion and cross-scale perception[J]. Journal on Communications, 2025, 46 (1): 108- 123
|
|
|
| [18] |
NARAYANAN M. SENetV2: aggregated dense layer for channelwise and global representations [EB/OL]. [2025-09-10]. https://arxiv.org/abs/2311.10807.
|
|
|
| [19] |
MAZUROV M. Railroad worker detection dataset [DB/OL]. [2025-05-15]. https://www.kaggle.com/datasets/johnsmith/shop-ping-trends.
|
|
|
| [20] |
TAN M, PANG R, LE Q V. EfficientDet: scalable and efficient object detection [C]//Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition. Seattle: IEEE, 2020: 10778–10787.
|
|
|
| [21] |
YANG G, LEI J, ZHU Z, et al. AFPN: asymptotic feature pyramid network for object detection [C]//Proceedings of the IEEE International Conference on Systems, Man, and Cybernetics. Honolulu: IEEE, 2024: 2184–2189.
|
|
|
| [22] |
JIANG Y, TAN Z, WANG J, et al. GiraffeDet: a heavy-neck paradigm for object detection [EB/OL]. [2025-09-10]. https://arxiv.org/abs/2202.04256.
|
|
|
| [23] |
DO NASCIMENTO M G, PRISACARIU V, FAWCETT R. DSConv: efficient convolution operator [C]//Proceedings of the IEEE/CVF International Conference on Computer Vision. Seoul: IEEE, 2020: 5147–5156.
|
|
|
| [24] |
HAN Q, FAN Z, DAI Q, et al. On the connection between local attention and dynamic depth-wise convolution [EB/OL]. [2025-09-10]. https://arxiv.org/abs/2106.04263.
|
|
|
| [25] |
CAO J, BAO W, SHANG H, et al GCL-YOLO: a GhostConv-based lightweight YOLO network for UAV small object detection[J]. Remote Sensing, 2023, 15 (20): 4932
doi: 10.3390/rs15204932
|
|
|
| [26] |
LI C, ZHOU A, YAO A. Omni-dimensional dynamic convolution [EB/OL]. [2025-09-10]. https://arxiv.org/abs/2209.07947.
|
|
|
| [27] |
WANG Z, LI C, XU H, et al. Mamba YOLO: a simple baseline for object detection with state space model [EB/OL]. [2025-09-10]. https://arxiv.org/abs/2406.05835.
|
|
|
| [28] |
TANG Y, HAN K, GUO J, et al GhostNetv2: enhance cheap operation with long-range attention[J]. Advances in Neural Information Processing Systems, 2022, 35: 9969- 9982
doi: 10.52202/068431-0724
|
|
|
| [29] |
CHENG T, SONG L, GE Y, et al. YOLO-world: real-time open-vocabulary object detection [C]//Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition. Seattle: IEEE, 2024: 16901–16911.
|
|
|
|
Viewed |
|
|
|
Full text
|
|
|
|
|
Abstract
|
|
|
|
|
Cited |
|
|
|
|
| |
Shared |
|
|
|
|
| |
Discussed |
|
|
|
|
