Research on a Lightweight Vision-based Monitoring Method for Rigid Guide Rail Joint Defects in Coal Mine Shafts
DOI:
https://doi.org/10.54097/2fhemd23Keywords:
Rigid guide rails, gap measurement, YOLOv10n, edge extraction, lightweight modelAbstract
Prolonged mechanical impacts on mine shaft rigid guide rails inevitably induce joint gaps and misalignments. However, existing visual inspection methods struggle to strike an optimal balance among detection accuracy, inference speed, and edge-deployability. To address these challenges, we propose YOLOv10n-LBC, a highly efficient, integrated model tailored for defect detection, gap measurement, and rail edge extraction on resource-constrained edge devices. Specifically, the network incorporates a Lightweight Ghost Convolutional CSP (LGC-CSP) module to compress the architecture, a Cross-Scale Decoupled (CSD) head to enhance multi-scale feature fusion, and a BiFormer attention mechanism to compensate for feature degradation caused by the lightweight design. Following detection, joint gap widths are quantified using a pixel-to-physical ratio methodology, while misalignments are identified via depth estimation-based edge extraction. Extensive experiments on a custom 6,100-image dataset demonstrate that YOLOv10n-LBC achieves a Precision of 98.6%, a Recall of 97.1%, and an mAP@0.5:0.95 of 72.2%. Despite its robust performance, the model maintains a highly compact footprint with only 1.57 M parameters and a 3.4 MB weight file. Real-world mine field tests validate the system's stability, maintaining maximum gap measurement errors below 1 mm, thereby offering a reliable solution for intelligent underground shaft inspection.
Downloads
References
[1] WANG Guofa, PANG Yihui, REN Huaiwei, et al. System engineering and key technologies research and practice of smart mine [J]. Journal of China Coal Society, 2024, 49(1):181−202.
[2] DU Fei, MA Tianbing, HU Weikang, et al. Fault diagnosis of rigid guide based on wavelet transform and improved convolutional neural network [J]. Industry and Mine Automation, 2022, 48(09):42-48+62.
[3] MA Tianbing, WANG Xinquan, WANG Xiaodong. Fault diagnosis method for rigid tank channel based on EMD-PNN network [J]. Journal of Electronic Measurement and Instrumentation, 2019, 33(03):58-64.
[4] ZHANG Yi, KONG Xianggao, GUO Mengda. Research and application of optical fiber sense technology in auxiliary shaft hoisting system [J]. Coal Engineering, 2025, 57(09):175-182.
[5] YANG Wanlin. Research on Wellbore Safety Detection Based on Inertial and 3D Laser Technology [D]. Tai’an:Shandong University of Science and Technology, 2020,12-25.
[6] NIU Weifeng, ZONG Liangliang, WU Feng, et al. Two-dimensional laser scanner based coal mine shaft guide deformation detection device [J]. Industry and Mine Automation, 2021, 47(03):101-104+111.
[7] Pang Meng. Research on Wellbore Inspection System of Deep Vertical Shaft Based on Machine Vision [D]. Huai’nan: Anhui University of Science and Technology, 2023, 35-54.
[8] ZHAO Baiting, WU Jundong, JIA Xiaofang. Lightweight guide defect detection algorithm based on feature enhancement [J]. Journal of Electronic Measurement and Instrumentation, 2023, 37(06):159-168.
[9] Sun Zhicheng. Rigid Tank Channel Defect Detection Based on Improved YOLOv5 and Embedded Platform[D]. Huai’nan: Anhui University of Science and Technology, 2024, 27-48.
[10] Fang Jiaxin. Research on Fault Detection of Guiding Device in Vertical Shaft Hoisting System Based on Machine Vision [D]. Huai’nan: Anhui University of Science and Technology, 2024, 41-58.
[11] WANG Manli, YANG Shuang, ZHANG Changsen. An improved YOLOv8n based dislocation detection algorithm for shaft rigid tank channel joint [J]. Coal Science and Technology, 2024, 52(S2):236-248.
[12] 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 (CVPR). Piscataway, NJ: IEEE, 2016: 779-788.
[13] Redmon J, Farhadi A. YOLO9000: Better, faster, stronger [C]// Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (CVPR). Piscataway, NJ: IEEE, 2017: 7263-7271.
[14] Redmon J, Farhadi A. YOLOv3: An incremental improvement [EB/OL]. arXiv preprint, arXiv:1804.02767, 2018.
[15] Bochkovskiy A, Wang C Y, Liao H Y M. YOLOv4: Optimal speed and accuracy of object detection [EB/OL]. arXiv preprint, arXiv:2004.10934, 2020.
[16] WANG C, YEH I, LIAO H. YOLOv9: Learning What You Want to Learn Using Programmable Gradient Information [EB/OL]. arXiv preprint, arXiv:2402.13616, 2024.
[17] Wang A, Chen H, Liu L, et al. YOLOv10: Real-time End-to-End Object Detection [EB/OL]. arXiv preprint arXiv: 2405.14458, 2024.
[18] Zhu L, Wang X, Ke Z, et al. BiFormer: Vision Transformer with Bi-Level Routing Attention [C]//Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR). 2023: 10323-10333.
[19] Wang Chien-Yao, Liao H Y M, Wu Y H, et al. CSPNet: A new backbone that can enhance learning capability of CNN [C]// Proceedings of the IEEE/CVF Conf on Computer Vision and Pattern Recognition Workshops. Piscataway, NJ: IEEE, 2020: 390-391.
[20] Han Kai, Wang Yunhe, Tian Qi, et al. Ghostnet: More features from cheap operations [C]// Proceedings of the IEEE/CVF Conf on Computer Vision and Pattern Recognition. Piscataway, NJ: IEEE, 2020: 1580-1589.
[21] Tang Yehui, Han Kai, Guo Jianyuan, et al. GhostNetv2: Enhance cheap operation with long-range attention [C]//Proc of the 36th Advances in Neural Information Processing Systems. Cambridge, MA: MIT Press, 2022: 9969-9982.
[22] Soudy M, Afify Y, Badr N. RepConv: A novel architecture for image scene classification on Intel scenes dataset [J]. International Journal of Intelligent Computing and Information Sciences, 22(2):63-73, 2022.
Downloads
Published
Issue
Section
License
Copyright (c) 2026 International Journal of Advanced Engineering and Technology Research
This work is licensed under a Creative Commons Attribution-NonCommercial 4.0 International License.
