The traction/braking performance of rail transit vehicles depends on the wheel rail contact condition. However the running performance of vehicles is quite different under different rail surface conditions, such as dry, wet and greasy. In view of the problems of large subjectivity and serious lag in traditional artificial experience based rail surface state recognition, a rail surface state recognition method based on an improved ResNet-50 deep learning network is proposed. Firstly, the ResNet-50 network is used to build the rail surface state recognition model. Secondly, transfer learning is introduced to improve the structure and parameters of the ResNet-50 network, and the improved ResNet-50 network is used to train the image data of the rail surface state. Finally, a model that can be used for classification is obtained to recognize the rail surface state. The results show that the optimized ResNet-50 model is more effective in identifying the rail surface state compared with the traditional ResNet-50 model. The model performance is better, and the identification accuracy can reach 92.75%.
References
[1]
Liu, J., Liu, L., He, J., Zhang, C. and Zhao, K. (2020) Wheel/Rail Adhesion State Identification of Heavy-Haul Locomotive Based on Particle Swarm Optimization and Kernel Extreme Learning Machine. JournalofAdvancedTransportation, 2020, Article ID: 8136939. https://doi.org/10.1155/2020/8136939
[2]
Zou, R., Ma, W. and Luo, S. (2018) Influence of the Wheel Diameter Difference on the Wheel/Rail Dynamic Contact Relationship of the Heavy Haul Locomotive. AustralianJournalofMechanicalEngineering, 16, 98-108. https://doi.org/10.1080/14484846.2018.1456787
[3]
Pichlik, P. and Bauer, J. (2021) Adhesion Characteristic Slope Estimation for Wheel Slip Control Purpose Based on UKF. IEEETransactionsonVehicularTechnology, 70, 4303-4311. https://doi.org/10.1109/tvt.2021.3072484
[4]
Wu, B., Wu, T., Wen, Z. and Jin, X. (2016) Numerical Analysis of High-Speed Wheel/Rail Adhesion under Interfacial Liquid Contamination Using an Elastic-Plastic Asperity Contact Model. ProceedingsoftheInstitutionofMechanicalEngineers, PartJ:JournalofEngineeringTribology, 231, 63-74. https://doi.org/10.1177/1350650116645025
[5]
He, J., Liu, G., Liu, J., Zhang, C. and Cheng, X. (2018) Identification of a Nonlinear Wheel/Rail Adhesion Model for Heavy-Duty Locomotives. IEEEAccess, 6, 50424-50432. https://doi.org/10.1109/access.2018.2868177
[6]
Chang, C., Chen, B., Cai, Y. and Wang, J. (2019) An Experimental Study of High Speed Wheel-Rail Adhesion Characteristics in Wet Condition on Full Scale Roller Rig. Wear, 440, Article 203092. https://doi.org/10.1016/j.wear.2019.203092
[7]
Liu, X., Xiao, C. and Meehan, P.A. (2019) The Effect of Rolling Speed on Lateral Adhesion at Wheel/Rail Interface under Dry and Wet Condition. Wear, 438, Article 203073. https://doi.org/10.1016/j.wear.2019.203073
[8]
Feng, H., Jiang, Z., Xie, F., Yang, P., Shi, J. and Chen, L. (2014) Automatic Fastener Classification and Defect Detection in Vision-Based Railway Inspection Systems. IEEETransactionsonInstrumentationandMeasurement, 63, 877-888. https://doi.org/10.1109/tim.2013.2283741
[9]
Torabi, M., Mousavi, S.G.M. and Younesian, D. (2018) A High Accuracy Imaging and Measurement System for Wheel Diameter Inspection of Railroad Vehicles. IEEETransactionsonIndustrialElectronics, 65, 8239-8249. https://doi.org/10.1109/tie.2018.2803780
[10]
Chandra, B.S., Sastry, C.S. and Jana, S. (2019) Robust Heartbeat Detection from Multimodal Data via CNN-Based Generalizable Information Fusion. IEEETransactionsonBiomedicalEngineering, 66, 710-717. https://doi.org/10.1109/tbme.2018.2854899
[11]
He, J., Yin, L., Liu, J., Zhang, C. and Yang, H. (2022) A Fault Diagnosis Method for Unbalanced Data Based on a Deep Cost Sensitive Convolutional Neural Network. IFAC-PapersOnLine, 55, 43-48. https://doi.org/10.1016/j.ifacol.2022.05.008
[12]
Liu, J., Yang, H., He, J., Sheng, Z. and Chen, S. (2022) Unbalanced Fault Diagnosis Based on an Invariant Temporal-Spatial Attention Fusion Network. ComputationalIntelligenceandNeuroscience, 2022, Article ID: 1875011. https://doi.org/10.1155/2022/1875011
[13]
Yu, Y., Sun, W., Liu, J. and Zhang, C. (2022) Traffic Flow Prediction Based on Depthwise Separable Convolution Fusion Network. JournalofBigData, 9, Article No. 83. https://doi.org/10.1186/s40537-022-00637-9
[14]
Xiong, Z., Li, Q., Mao, Q. and Zou, Q. (2017) A 3D Laser Profiling System for Rail Surface Defect Detection. Sensors, 17, Article 1791. https://doi.org/10.3390/s17081791
[15]
Yu, H., Li, Q., Tan, Y., Gan, J., Wang, J., Geng, Y., et al. (2019) A Coarse-to-Fine Model for Rail Surface Defect Detection. IEEETransactionsonInstrumentationandMeasurement, 68, 656-666. https://doi.org/10.1109/tim.2018.2853958
[16]
Li, Q. and Ren, S. (2012) A Visual Detection System for Rail Surface Defects. IEEETransactionsonSystems, Man, andCybernetics, PartC (ApplicationsandReviews), 42, 1531-1542. https://doi.org/10.1109/tsmcc.2012.2198814
[17]
Zhang, H., Jin, X., Wu, Q.M.J., Wang, Y., He, Z. and Yang, Y. (2018) Automatic Visual Detection System of Railway Surface Defects with Curvature Filter and Improved Gaussian Mixture Model. IEEETransactionsonInstrumentationandMeasurement, 67, 1593-1608. https://doi.org/10.1109/tim.2018.2803830
[18]
Hu, Z., Zhu, H., Hu, M. and Ma, Y. (2018) Rail Surface Spalling Detection Based on Visual Saliency. IEEJTransactionsonElectricalandElectronicEngineering, 13, 505-509. https://doi.org/10.1002/tee.22594
[19]
Gan, J., Wang, J., Yu, H., Li, Q. and Shi, Z. (2020) Online Rail Surface Inspection Utilizing Spatial Consistency and Continuity. IEEETransactionsonSystems, Man, andCybernetics:Systems, 50, 2741-2751. https://doi.org/10.1109/tsmc.2018.2827937
[20]
Gibert, X., Patel, V.M. and Chellappa, R. (2017) Deep Multitask Learning for Railway Track Inspection. IEEETransactionsonIntelligentTransportationSystems, 18, 153-164. https://doi.org/10.1109/tits.2016.2568758
