Research on lightweight rail fastener inspection model based on YOLOv5

XIE Xing; WANG Yue; LI Liming; ZHENG Shubin; PENG Lele; ZHU Ting

doi:10.12299/jsues.24-0050

Volume 39 Issue 3

Sep. 2025

Turn off MathJax

Article Contents

Article Navigation > Journal of Shanghai University of Engineering Science > 2025 > 39(3): 257-265

XIE Xing, WANG Yue, LI Liming, ZHENG Shubin, PENG Lele, ZHU Ting. Research on lightweight rail fastener inspection model based on YOLOv5[J]. Journal of Shanghai University of Engineering Science, 2025, 39(3): 257-265. doi: 10.12299/jsues.24-0050

Citation:

XIE Xing, WANG Yue, LI Liming, ZHENG Shubin, PENG Lele, ZHU Ting. Research on lightweight rail fastener inspection model based on YOLOv5[J]. Journal of Shanghai University of Engineering Science, 2025, 39(3): 257-265. doi: 10.12299/jsues.24-0050

Citation:

PDF( 2655 KB)

Research on lightweight rail fastener inspection model based on YOLOv5

doi: 10.12299/jsues.24-0050

XIE Xing^{1a
,},
WANG Yue^1a,
LI Liming^{1a, 1b
,
,},
ZHENG Shubin^1b,
PENG Lele^{1a, 1b},
ZHU Ting²

1a.
School of Urban Railway Transportation
1b.
Shanghai Rail Transit Vibration and Noise Control Technology Engineering Research Center, Shanghai University of Engineering Science, Shanghai 201620, China
2.
China Railway Shanghai Bureau Group Co., Ltd. Science and Technology Research Institute, Shanghai, 200070

Received Date: 2024-03-04
Available Online: 2025-12-22
Publish Date: 2025-09-30

Abstract

Abstract

Aiming at the issues of model parameter size, detection accuracy, and efficiency in rail fastener detection algorithms, a lightweight network model named YOLO-FBS based on YOLOv5n was proposed. The model embeds the Faster Block module to achieve a reduction in network parameters, introduces the SimAM parameter-free feature selection module and the CAFARE lightweight upsampling module to improve detection accuracy, and finally employs LAMP score-based pruning to remove channels with lower weights. The resulting YOLO-FBS model has only 0.28×10⁶ parameters, achieves a detection accuracy of 89.2%, and reaches a detection speed of 190.2 frames/s. This network model offers the advantages of low parameter count and high detection speed while maintaining high detection accuracy.
- YOLOv5,
- lightweight,
- SimAM,
- pruning,
- fastener detection

FullText(HTML)

References(15)

References

[1]	李海锋, 许玉德. 计算机编制铁路轨道养护维修计划的方法[J] . 同济大学学报, 2004, 32(4): 480 − 484.
[2]	张曦. 浅析超声波探伤技术在钢轨探伤中的应用[J] . 中国设备工程, 2023(5): 130 − 132. doi: 10.3969/j.issn.1671-0711.2023.05.054
[3]	叶杭璐, 王超, 吴杨娜, 等. 轨道交通涡流探伤仪的设计与实现[J] . 浙江树人大学学报(自然科学版), 2016, 16(3): 7 − 11.
[4]	韦若禹, 李舒婷, 吴松荣, 等. 基于改进YOLO V3算法的轨道扣件缺陷检测[J] . 铁道标准设计, 2020, 64(12): 30 − 36.
[5]	邹文武, 许贵阳, 白堂博. 基于EfficientDet的轨道扣件识别与检测[J] . 武汉大学学报(工学版), 2024, 57(7): 1006 − 1012.
[6]	HOWARD A G, ZHU M L, CHEN B, et al. MobileNets: efficient convolutional neural networks for mobile vision applications[EB/OL] . (2017-04-17)[2021-01-12] . https://arxiv.org/abs/1704.04861.
[7]	SANDLER M, HOWARD A, ZHU M L, et al. MobileNetV2: inverted residuals and linear bottlenecks[C] //Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition. Salt Lake City: IEEE, 2018: 4510-4520.
[8]	ZHANG X Y, ZHOU X Y, LIN M X, et al. ShuffleNet: an extremely efficient convolutional neural network for mobile devices[C] //Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition. Salt Lake City: IEEE, 2018: 6848−6856.
[9]	HAN K, WANG Y H, TIAN Q, et al. GhostNet: more features from cheap operations[C] //Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition. Seattle: IEEE, 2020: 1577−1586.
[10]	CHEN J R, 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.
[11]	YANG L X, ZHANG R Y, LI L D, et al. SimAM: a simple, parameter-free attention module for convolutional neural networks[J] . PMLR, 2021, 139: 11863−11874.
[12]	WANG J Q, CHEN K, XU R, et al. CARAFE: content-aware ReAssembly of FEatures[C] //IEEE/CVF International Conference on Computer Vision (ICCV). Seoul: IEEE, 2019: 3007−3016.
[13]	LEE J, PARK S, MO S, et al. Layer-adaptive sparsity for the magnitude-based pruning[EB/OL] . (2020-10-15)[2024-01-15] . https://doi.org/10.48550/arXiv.2010.07611.
[14]	HU J, SHEN L, SUN G. Squeeze-and-excitation networks[C] //IEEE/CVF Conference on Computer Vision and Pattern Recognition. Salt Lake City: IEEE, 2018: 7132−7141.
[15]	WOO S, PARK J, LEE J Y, et al. CBAM: convolutional block attention module[C] //Proceedings of the 15th European Conference on Computer Vision. Munich: Springer, 2018: 3−19.