Automatyczna diagnostyka elementów toru kolejowego z użyciem sieci neuronowych głębokiego uczenia

pol Artykuł w języku polskim DOI: 10.14313/PAR_253/21

wyślij Piotr Bojarczak , Waldemar Nowakowski Uniwersytet Radomski im. Kazimierza Pułaskiego, ul. Malczewskiego 29, 26-600 Radom

Pobierz Artykuł

Streszczenie

W artykule omówiono metodę automatycznej diagnostyki szlaku kolejowego. Polega ona na automatycznej ocenie stanu technicznego wybranych elementów toru, takich jak szyny, podkłady drewniane, podkłady betonowe, przytwierdzenia podkładów oraz rozjazdy. Przeprowadzono ją na podstawie analizy obrazów wideo elementów toru kolejowego zarejestrowanych przez dwie kamery linijkowe umieszczone na drezynie pomiarowej. Do oceny stanu technicznego badanych elementów zastosowano wybraną sieć neuronową głębokiego uczenia FCN-8. Określono też skuteczność zastosowanego algorytmu na podstawie takich miar jak IoU, Precision, Recall. Przedstawiono wnioski dotyczące zastosowania sieci FCN-8 w automatycznej klasyfikacji cech wybranych elementów toru kolejowego. Uzyskane rezultaty porównano z innymi metodami wykorzystywanymi w diagnostyce wizyjnej.

Słowa kluczowe

algorytmy uczące, diagnostyka wizyjna, przetwarzanie obrazów, tor kolejowy

Automatic Diagnosis of Railway Track Elements Using Deep Learning Neural Networks

Abstract

The article discusses a method for automatic diagnostics of a railway track. It consists in automatic evaluation of the technical condition of selected track elements, such as rails, wooden and concrete sleepers, fasteners and turnouts. It was carried out on the basis of analysis of video images of railroad track elements recorded by two line cameras placed on the diagnostic carriage. The selected FCN-8 deep learning neural network was used to assess the technical condition of the surveyed elements, and the effectiveness of the applied algorithm was determined on the basis of such measures as IoU, Precision, Recall. Conclusions on the application of the FCN-8 network in the automatic classification of features of selected railroad track elements are presented. The results obtained were compared with other methods used in vision diagnostics.

Keywords

image processing, learning algorithm, railway track, vision diagnostic

Bibliografia

  1. Babenko P., Visual inspection of railroad tracks. PhD thesis, University of Central Florida, 2009.
  2. Bojarczak P., Application of wavelet transform into precise localization of railway rail edges in visual diagnostic of track, “Archives of Transport”, Vol. 24, No. 1, 2012, 5–16, DOI: 10.2478/v10174-012-0001-9.
  3. Bojarczak P., Visual algorithms for automatic detection of squat flaws in railway rails, “Insight – Non-Destructive Testing and Condition Monitoring”, Vol. 55, No. 7, 2013, 353–359, DOI: 10.1784/insi.2012.55.7.353.
  4. Bojarczak P., Lesiak P., Application of neural networks into automatic visual diagnostic of railway wooden sleepers, Międzynarodowa Konferencja Naukowa Transport XXI w., Białowieża 2010, “Logistyka”, Nr 4, 2010.
  5. Bojarczak P., Lesiak P., SVM based classification method of railway’s defects, “Pomiary Automatyka Kontrola”, R. 53, Nr 12, 2007, 15–17.
  6. Bojarczak P., Lesiak P., Visual system diagnosing the state of elements fastening the rail to the sleepers. “Pomiary Automatyka Kontrola”, R. 57, Nr 12, 2011, 1605–1607.
  7. Chen L.C., Papandreou G., Kokkinos I., Murphy K., Yuille A.L., DeepLab: Semantic Image Segmentation with Deep Convolutional Nets, Atrous Convolution and Fully Connected CRFs, “Computer Vision and Pattern Recognition”, 2018, DOI: 10.48550/arXiv.1606.00915.
  8. Chen L.C., Papandreou G., Kokkinos I., Murphy K., Yuille A.L., Semantic Image Segmentation with Deep Convolutional Nets and Fully Connected CRFs, “Computer Vision and Pattern Recognition”, 2014, DOI: 10.48550/arXiv.1412.7062.
  9. Chen L.C., Papandreou G., Schroff F., Adam H., Rethinking Atrous Convolution for Semantic Segmentation, “Computer Vision and Pattern Recognition”, 2017, DOI: 10.48550/arXiv.1706.05587.
  10. Chen L.C., Papandreou G., Schroff F., Adam H., Encoder-Decoder with Atrous Separable Convolution for Semantic Image Segmentation. “Computer Vision and Pattern Recognition”, 2018, DOI: 10.48550/arXiv.1802.02611.
  11. Camargo L., Edwards J., Barkan C., Emerging condition monitoring technologies for railway track components and special track work. ASME/ASCE/IEEE Joint Rail Conference, Pueblo, Colorado, 2011, DOI: 10.1115/JRC2011-56113.
  12. Dai P., Du X., Wang S., Gu Z., Ma Y., Rail fastener automatic method in complex background. Proceedings of the SPIE, Vol. 10806, 2018, DOI: 10.1117/12.2503323.
  13. Dumoulin V., Visin F., A guide to convolution arithmetic for deep learning, “Machine Learning”, 2016, DOI: 10.48550/arXiv.1603.07285.
  14. Feng H., Jiang Z., Xie F., Yang P., Shi J., Chen L., Automatic fastener classification and defect detection in vision based railway inspection systems, “IEEE Transactions on Instrumentation and Measurement”, Vol. 63, No. 4, 2014, 877–888, DOI: 10.1109/TIM.2013.2283741.
  15. Gibert X., Patel V.M., Chellappa R., Deep Multitask Learning for Railway Track Inspection, “IEEE Transactions on Intelligent Transportation Systems”, Vol. 18, No. 1, 2017, 153–164, DOI: 10.1109/TITS.2016.2568758.
  16. Giben X., Patel V.M., Chellappa R., Material classification and semantic segmentation of railway track images with deep convolutional neural networks. IEEE International Conference on Image Processing, 2015, 621–625, DOI: 10.1109/ICIP.2015.7350873.
  17. Gibert X., Patel V.M., Chellappa R., Robust fastener detection for autonomous visual railway track inspection. IEEE Winter Conference on Application of Computer Vision, 2015, 694–701, DOI: 10.1109/WACV.2015.98.
  18. He K., Gkioxari G., Dollar P., Girshick R., Mask R-CNN, “Computer Vision and Pattern Recognition”, 2017, DOI: 10.48550/arXiv.1703.06870.
  19. Karakose M., Yaman O., Murat K., Akin E., A new approach for condition monitoring and detection of rail components and rail track in railway, “International Journal of Computational Intelligence Systems”, Vol. 11, No. 1, 2018, 830–845, DOI: 10.2991/ijcis.11.1.63.
  20. Karakose M., Yaman O., Murat K., Akin E., Real time implementation for fault diagnosis and condition monitoring approach using image processing in railway switches. “International Journal of Applied Methods in Electronics and Computers”, 2016, 307–313, DOI: 10.18100/ijamec.270627.
  21. Krizhevsky A., Sutskever I., Hinton G.E., ImageNet classification with Deep Convolutional Neural Networks, “Advances in Neural Information Processing Systems”, Vol. 14, 2012, 1097–1105.
  22. Kumar B.V.K.V., Mahalanobis A., Juday R., Correlation pattern recognition, Cambridge University Press, 2005.
  23. Lesiak P., Mobilna diagnostyka szyn w torze kolejowym. Monografia habilitacyjna, Wydział Transportu Politechniki Warszawskiej, 2008.
  24. Lesiak P., Bojarczak P., Application of neural classifier to railway flaw detection in the method of metal magnetic memory, The 6 International Conference “Environmental Engineering” Selected papers, Vol. 2, 744–747, Vilnius, Lithuania, 2005.
  25. Lesiak P., Bojarczak P., Application of wavelets and fuzzy sets to detection of head – checking defects in railway rails, “Transport Systems Telematics”, 10 Conference TST 2010, Communications in Computer and Information Science, Springer, Vol. 104, 2010, 327–334, DOI: 10.1007/978-3-642-16472-9_36.
  26. Lesiak P., Bojarczak P., Inteligentne algorytmy analizy ultradźwiękowej obrazów w badaniach bezstykowych złączy szyn metodą TOFD, „Logistyka”, 3, 2012.
  27. Lesiak P., Bojarczak P., Migdal M., Inteligentne klasyfikatory wad kontaktowo-naprężeniowych w szynach kolejowych. „Pomiary Automatyka Komputery w Gospodarce i Ochronie Środowiska”, Nr 3, 2009, 13–17.
  28. Li Y., Trinh H., Haas N., Otto C., Pankanti S., Rail component detection, optimization, and assessment for automatic rail track inspection, “IEEE Transactions on Intelligent Transportation Systems”, Vol. 15, No. 2, 2014, 760–770,
  29. Liu W., Anguelov D., Erhan D., Szegedy C., Reed S., Fu C.Y., Berg A.C., SSD: Single Shot MultiBox Detector, “Computer Vision and Pattern Recognition”, 2015, DOI: 10.1007/978-3-319-46448-0_2.
  30. Long J., Shelhamer E., Darrell T., Fully Convolutional Networks for Semantic Segmentation, “Computer Vision and Pattern Recognition”, 2014, DOI: 10.48550/arXiv.1411.4038.
  31. Mazzeo P., Nitti M., Stella E., Distante A., Visual recognition of fastening bolts for railroad maintenance, “Pattern Recognition Letters”, Vol. 25, No. 6, 2004, 669–777, DOI: 10.1016/j.patrec.2004.01.008.
  32. Rauschmayr N., Hoechemer M., Zurkirchen M., Kenzelmann S., Gilles M., Deep Learning of Railway Track Faults using GPUs, Global Technology Conference, Santa Clara, USA, 2018.
  33. Redmon J., Divvala S., Girshick R., Farhadi A., You Only Look Once: Unified, Real-Time Object Detection, Proceedings of IEEE Conference on Computer Vision and Pattern Recognition, 2016, 779–788, DOI: 10.1109/CVPR.2016.91.
  34. Ren S., He K., Girshick R., Sun J., Faster R-CNN: Towards real-time object detection with region proposal networks, “IEEE Transactions on Pattern Recognition and Machine Intelligence”, Vol. 39, 2017, 1137–1149, DOI: 10.1109/TPAMI.2016.2577031.
  35. Resendiz E., Hart J., Ahuja N., Automated visual inspec tion of railroad tracks, “IEEE Transactions on Intelligent Transportation Systems”, Vol. 14, No. 2, 2013, 751–760, DOI: 10.1109/TITS.2012.2236555.
  36. Tan C., Sun F., Kong T., Zhang W., Yang C., Liu C., A survey on Deep Transfer Learning, Proceedings of 27th International Conference on Artificial Neural Networks, Rhodes, part 3, Vol. 11141, 2018, 270–279, DOI: 10.1007/978-3-030-01424-7_27.
  37. Yang J., Tao W., Liu M., Zhang Y., Zang H., Zhao H., An efficient direction field-based method for the detection of fasteners on high-speed railways, “Sensors”, Vol. 11, No. 8, 2011, 364–7381, DOI: 10.3390/s110807364.
  38. Shorten C., Khoshgoftaar T.M., A survey on Image Data Augmentation for Deep Learning. “Journal of Big Data”, Vol. 6, 2019, DOI: 10.1186/s40537-019-0197-0