|
|
基于改进YOLOv8n的钢轨表面伤损检测算法
|
Abstract:
在采集钢轨表面伤损数据建立数据集过程中，发现钢轨表面伤损存在大尺度伤损较多及尺度变化较大的情况，针对这种情况提出了一种面向多尺度钢轨表面伤损的改进YOLOv8n检测算法。首先，将网络中的C2f与Triplet Attention融合设计新的特征提取模块C2f-TA，扩大网络层特征图的感受野，增加对大尺度目标的检测精度。然后，结合网络层连接模块SDI与TFE，设计新的多尺度特征融合Neck结构，通过在不同的网络层将Concat模块替换为TFE模块和SDI模块，实现对四个不同尺度级别的特征图的融合，赋予特征图丰富的语义特征和细节特征；同时，使用SSFF模块和CPAM模块，构建新的小尺度目标检测分支，利用通道和局部注意力机制，增强小尺度目标的特征表达。结果表明，较基准算法Precision提升了1.8%，Recall提升了7.2%，mAP@50和mAP@50-95分别提升了3%和2.8%。
In the process of collecting rail surface damage data and establishing dataset, it is found that there are many large-scale damages and large-scale change of rail surface damage. In response to this, an improved YOLOv8n detection algorithm tailored for multi-scale Rail surface damage was proposed. Firstly, a new feature extraction module, C2f-TA, was designed by integrating C2f with Triplet Attention in the network to expand the receptive field of the feature maps, thereby enhancing the detection accuracy for large-scale targets. Subsequently, a novel multi-scale feature fusion Neck structure was devised by combining the network layer connection modules SDI and TFE. This was achieved by replacing the Concat modules with TFE and SDI modules at different network layers, enabling the fusion of feature maps across four distinct scale levels, thereby enriching the semantic and detailed features of the feature maps. Additionally, a new small-scale target detection branch was constructed using the SSFF and CPAM modules, leveraging channel and local attention mechanisms to strengthen the feature representation of small-scale targets. The results demonstrate that compared to the baseline algorithm, the Precision is improved by 1.8%, the Recall by 7.2%, and the mAP@50 and mAP@50-95 are enhanced by 3% and 2.8%, respectively.
| [1] | Cannon, D.F., Edel, K.O., Grassie, S.L. and Sawley, K. (2003) Rail Defects: An Overview. Fatigue & Fracture of Engineering Materials & Structures, 26, 865-886. https://doi.org/10.1046/j.1460-2695.2003.00693.x |
| [2] | Ph Papaelias, M., Roberts, C. and Davis, C.L. (2008) A Review on Non-Destructive Evaluation of Rails: State-Of-The-Art and Future Development. Proceedings of the Institution of Mechanical Engineers, Part F: Journal of Rail and Rapid Transit, 222, 367-384. https://doi.org/10.1243/09544097jrrt209 |
| [3] | Rajamäki, J., Vippola, M., Nurmikolu, A. and Viitala, T. (2016) Limitations of Eddy Current Inspection in Railway Rail Evaluation. Proceedings of the Institution of Mechanical Engineers, Part F: Journal of Rail and Rapid Transit, 232, 121-129. https://doi.org/10.1177/0954409716657848 |
| [4] | Kishore, M.B., Park, J.W., Song, S.J., Kim, H.J. and Kwon, S.G. (2019) Characterization of Defects on Rail Surface Using Eddy Current Technique. Journal of Mechanical Science and Technology, 33, 4209-4215. https://doi.org/10.1007/s12206-019-0816-x |
| [5] | Kang, D., Oh, J., Kim, J. and Park, S. (2015) Study on MFL Technology for Defect Detection of Railroad Track under Speed-Up Condition. Journal of the Korean Society for Railway, 18, 401-409. https://doi.org/10.7782/jksr.2015.18.5.401 |
| [6] | Antipov, A.G. and Markov, A.A. (2019) Detectability of Rail Defects by Magnetic Flux Leakage Method. Russian Journal of Nondestructive Testing, 55, 277-285. https://doi.org/10.1134/s1061830919040028 |
| [7] | Xue, Z., Xu, Y., Hu, M. and Li, S. (2023) Systematic Review: Ultrasonic Technology for Detecting Rail Defects. Construction and Building Materials, 368, Article ID: 130409. https://doi.org/10.1016/j.conbuildmat.2023.130409 |
| [8] | Li, Y.J., Yao, F.T., Jiao, S.B., Huang, W.C. and Zhang, Q. (2020) Identification and Classification of Rail Damage Based on Ultrasonic Echo Signals. 2020 39th Chinese Control Conference (CCC), Shenyang, 27-29 July 2020, 3077-3082. https://doi.org/10.23919/ccc50068.2020.9189607 |
| [9] | Kundu, T., Datta, A.K., Topdar, P. and Sengupta, S. (2024) Optimal Location of Acoustic Emission Sensors for Detecting Rail Damage. Proceedings of the Institution of Civil Engineers—Structures and Buildings, 177, 254-263. https://doi.org/10.1680/jstbu.21.00074 |
| [10] | Bruzelius, K. and Mba, D. (2004) An Initial Investigation on the Potential Applicability of Acoustic Emission to Rail Track Fault Detection. NDT & E International, 37, 507-516. https://doi.org/10.1016/j.ndteint.2004.02.001 |
| [11] | Bai, T., Gao, J., Yang, J. and Yao, D. (2021) A Study on Railway Surface Defects Detection Based on Machine Vision. Entropy, 23, Article 1437. https://doi.org/10.3390/e23111437 |
| [12] | Li, Q. and Ren, S. (2012) A Visual Detection System for Rail Surface Defects. IEEE Transactions on Systems, Man, and Cybernetics, Part C (Applications and Reviews), 42, 1531-1542. https://doi.org/10.1109/tsmcc.2012.2198814 |
| [13] | Min, Y., Xiao, B., Dang, J., Yue, B. and Cheng, T. (2018) Real Time Detection System for Rail Surface Defects Based on Machine Vision. EURASIP Journal on Image and Video Processing, 2018, Article No. 3. https://doi.org/10.1186/s13640-017-0241-y |
| [14] | He, Z., Wang, Y., Yin, F. and Liu, J. (2016) Surface Defect Detection for High-Speed Rails Using an Inverse P-M Diffusion Model. Sensor Review, 36, 86-97. https://doi.org/10.1108/sr-03-2015-0039 |
| [15] | Choi, J. and Han, J. (2024) Deep Learning (Fast R-CNN)-Based Evaluation of Rail Surface Defects. Applied Sciences, 14, Article 1874. https://doi.org/10.3390/app14051874 |
| [16] | Guo, F., Qian, Y., Rizos, D., Suo, Z. and Chen, X. (2021) Automatic Rail Surface Defects Inspection Based on Mask R-CNN. Transportation Research Record: Journal of the Transportation Research Board, 2675, 655-668. https://doi.org/10.1177/03611981211019034 |
| [17] | Ren, S., He, K., Girshick, R. and Sun, J. (2017) Faster R-CNN: Towards Real-Time Object Detection with Region Proposal Networks. IEEE Transactions on Pattern Analysis and Machine Intelligence, 39, 1137-1149. https://doi.org/10.1109/tpami.2016.2577031 |
| [18] | He, K., Gkioxari, G., Dollar, P. and Girshick, R. (2017) Mask R-CNN. 2017 IEEE International Conference on Computer Vision (ICCV), Venice, 22-29 October 2017, 2980-2988. https://doi.org/10.1109/iccv.2017.322 |
| [19] | Wang, M. and Zhou, Y. (2024) Autonomous Rail Surface Defect Identification Based on an Improved One-Stage Object Detection Algorithm. Journal of Performance of Constructed Facilities, 38, Article ID: 04024041. https://doi.org/10.1061/jpcfev.cfeng-4816 |
| [20] | Arrosida, H., Susanto, I.A., Ciptaningrum, A., et al. (2023) Rail Line Surfaces Defect Monitoring using YOLO Architecture: Case Study on Madiun-Magetan Track, East Java. International Journal of Innovative Science and Research Technology, 8, 1150-1164. |
| [21] | Redmon, J., Divvala, S., Girshick, R. and Farhadi, A. (2016) You Only Look Once: Unified, Real-Time Object Detection. 2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR), Las Vegas, 27-30 June 2016, 779-788. https://doi.org/10.1109/cvpr.2016.91 |
| [22] | Misra, D., Nalamada, T., Arasanipalai, A.U. and Hou, Q. (2021) Rotate to Attend: Convolutional Triplet Attention Module. 2021 IEEE Winter Conference on Applications of Computer Vision (WACV), Waikoloa, 3-8 January 2021, 3138-3147. https://doi.org/10.1109/wacv48630.2021.00318 |
| [23] | Peng, Y., Sonka, M. and Chen, D.Z. (2023) U-Net v2: Rethinking the Skip Connections of U-Net for Medical Image Segmentation. arXiv: 2311.17791. |
| [24] | Kang, M., Ting, C., Ting, F.F. and Phan, R.C. (2024) ASF-YOLO: A Novel YOLO Model with Attentional Scale Sequence Fusion for Cell Instance Segmentation. Image and Vision Computing, 147, Article ID: 105057. https://doi.org/10.1016/j.imavis.2024.105057 |
| [25] | Lin, T., Dollar, P., Girshick, R., He, K., Hariharan, B. and Belongie, S. (2017) Feature Pyramid Networks for Object Detection. 2017 IEEE Conference on Computer Vision and Pattern Recognition (CVPR), Honolulu, 21-26 July 2017, 936-944. https://doi.org/10.1109/cvpr.2017.106 |
| [26] | Zheng, Y., Zhang, X., Zhang, R. and Wang, D. (2022) Gated Path Aggregation Feature Pyramid Network for Object Detection in Remote Sensing Images. Remote Sensing, 14, Article 4614. https://doi.org/10.3390/rs14184614 |
| [27] | Wang, Q., Gao, T., He, Q., Liu, Y., Wu, J. and Wang, P. (2022) Severe Rail Wear Detection with Rail Running Band Images. Computer-Aided Civil and Infrastructure Engineering, 38, 1162-1180. https://doi.org/10.1111/mice.12948 |
| [28] | Yue, B., Wang, Y., Min, Y., Zhang, Z., Wang, W. and Yong, J. (2019. Rail Surface Defect Recognition Method Based on Adaboost Multi-Classifier Combination. 2019 Asia-Pacific Signal and Information Processing Association Annual Summit and Conference (APSIPA ASC), Lanzhou, 18-21 November 2019, 391-396. https://doi.org/10.1109/apsipaasc47483.2019.9023096 |
| [29] | Yang, R., He, Y., Gao, B., Tian, G.Y. and Peng, J. (2015) Lateral Heat Conduction Based Eddy Current Thermography for Detection of Parallel Cracks and Rail Tread Oblique Cracks. Measurement, 66, 54-61. https://doi.org/10.1016/j.measurement.2015.01.024 |
| [30] | Zhai, W., Gao, J., Liu, P. and Wang, K. (2014) Reducing Rail Side Wear on Heavy-Haul Railway Curves Based on Wheel-Rail Dynamic Interaction. Vehicle System Dynamics, 52, 440-454. https://doi.org/10.1080/00423114.2014.906633 |
| [31] | Wang, C., Yeh, I. and Mark Liao, H. (2024) Yolov9: Learning What You Want to Learn Using Programmable Gradient Information. In: Leonardis, A., Ricci, E., Roth, S., Russakovsky, O., Sattler, T. and Varol, G., Eds., Computer Vision—ECCV 2024, Springer Nature Switzerland, 1-21. https://doi.org/10.1007/978-3-031-72751-1_1 |
| [32] | Wang, A., Chen, H., Liu, L., et al. (2024) Yolov10: Real-Time End-to-End Object Detection. arXiv: 2405.14458. |
| [33] | Koonce, B. (2021) ResNet 50. In: Koonce, B., Ed., Convolutional Neural Networks with Swift for Tensorflow, Apress, 63-72. https://doi.org/10.1007/978-1-4842-6168-2_6 |
| [34] | Sengupta, A., Ye, Y., Wang, R., Liu, C. and Roy, K. (2019) Going Deeper in Spiking Neural Networks: VGG and Residual Architectures. Frontiers in Neuroscience, 13, Article 95. https://doi.org/10.3389/fnins.2019.00095 |
