Intelligent Image Analysis Technology and Application for Rail Track Inspection

  • Peng Dai
  • Shengchun WangEmail author
  • Zichen Gu


Based on the deep learning technology, the track image intelligent recognition algorithm has been developed, which has made significant progress in the analysis of rail track infrastructure inspection. The main work includes as follows: (1) research on intelligent recognition method of fastener defect based on convolutional neural network—by modeling the deep network from the fasteners image of the nationwide rail line, the detection rate of abnormal fasteners has been significantly improved and the false alarm rate has been greatly reduced. The research achievement has been applied in China’s railway inspection. (2) Research on foreign object recognition algorithm in ballastless track based on heuristic deep learning—aiming at the problem of foreign objects falling in the roadbed of the ballastless track, the system has a single type of foreign object recognition capability by labeling and iterative training of the limited samples. Then, by adding the new foreign objects found in the model to the training set for iterative training, a heuristic learning framework is constructed, which makes the system recognize a variety of foreign objects and has been put on probation in China’s high-speed rail to verify the outstanding performance of the recognition algorithm. The research work shows that the development of artificial intelligence technology and industry integration will collide with a fierce spark, and the continuous absorption of new technologies and new methods is the way to promote technological innovation in the industry.


Deep learning Intelligent recognition Fastener defect Foreign object recognition 


  1. 1.
    Çaglar, A., Yousef, R., Dogru, S., et al.: Railway fastener inspection by real-time machine vision. IEEE Trans. Syst. Man Cybern. Syst. 45(7), 1101–1107 (2015)CrossRefGoogle Scholar
  2. 2.
    Xu, G., Shi, T., Ren, S., et al.: Development of on-board track inspection system based on computer vision. China Railway Sci. 34(1), 139–143 (2013)Google Scholar
  3. 3.
    Ren, S., Li, Q., Xu, G., et al.: Research on robust real-time rail surface scratch detection algorithm. China Railway Sci. 32(1), 25–29 (2011)Google Scholar
  4. 4.
    Li, Q., Zhang, H., Ren, S., et al.: Rail-wave grinding detection method based on frequency domain characteristics of rail images. China Railway Sci. 37(1), 24–30 (2016)Google Scholar
  5. 5.
    Zhou, W., Sun, Z., Ren, S., et al.: Measurement method of contact net geometry parameters based on multi-view stereo vision. China Railway Sci. 36(5), 104–109 (2015)Google Scholar
  6. 6.
    Maeeeo, P.L., Stella, E., Ancona, N., et al.: Visual detection of hexagonal headed bolts using method of frames and matching pursuit. In: The second Iberian conference on pattern recognition and image analysis, pp. 277–284. Estoril, Portugal (2005)Google Scholar
  7. 7.
    Marino, F., Distance, A., Mazzeo, P.L., et al.: A Real time visual inspection system for railway maintenance: automatic hexagonal headed bolts detection. IEEE Trans. Syst. Man Cybern. Part C Appl. Rev. 37(3), 418–428 (2007)CrossRefGoogle Scholar
  8. 8.
    Li, Y., Ttinh, H., Haas, N., et al.: Rail component detection, optimization, and assessment for automatic rail track inspection. IEEE Trans. Intell. Transp. Syst. 15(2), 760–770 (2014)CrossRefGoogle Scholar
  9. 9.
    Feng, H., Jiang, Z., Xie, F., et al.: Automatic fastener classification and defect detection in vision-based railway inspection systems. IEEE Trans. Instrum. Meas. 63(4), 877–888 (2014)CrossRefGoogle Scholar
  10. 10.
    Li, Y.: Research on State Detection Algorithm of Railway Line Based on Machine Vision. Southwest Jiaotong University, Chengdu (2013)Google Scholar
  11. 11.
    Li, Y., Li, B., Xiong, Y., et al.: State detection of railway fasteners based on hog characteristics. Sens. Microsyst. 32(10), 110–113 (2013)Google Scholar
  12. 12.
    Liu, J., Wang, K., Yuan, J., et al.: Optimization of RBF-SVM model in railway fastener image detection. Comput. Eng. Appl. 50(15), 30–33 (2015)Google Scholar
  13. 13.
    Liu, X., Mu, Y., Zhang, B.: Study on the detection algorithm of rail fastener based on computer vision. J. East China Jiaotong Univ. 34(2), 72–77 (2017)Google Scholar
  14. 14.
    Dou, Y., Huang, Y., Qingyong, L., et al.: A fast template matching-based algorithm for railway bolts detection. Int. J. Mach. Learn. Cybern. 5(6), 835–844 (2014)CrossRefGoogle Scholar
  15. 15.
    Gibert, X., Patel, V.M., Chellappa, R.: Material classification and semantic segmentation of railway track images with deep convolution neural networks. In: IEEE International Conference on Mobile Adhoc & Sensor Systems, pp. 92–101. Dallas, USA (2015)Google Scholar
  16. 16.
    Gibert, X., Patel, V.M., Chellappa, R.: Deep multi-task learning for railway track inspection. IEEE Trans. Intell. Transp. Syst. 18(1), 153–164 (2017)CrossRefGoogle Scholar
  17. 17.
    Du, X., Dai, P., Li, Y., et al.: Automatic detection algorithm for railway plug based on deep learning. China Railway Sci. 38(3), 89–96 (2017)Google Scholar
  18. 18.
    Xia, Y., Xie, F., Jiang, Z.: Broken railway fastener detection based on adaboost algorithm. In: International Conference on Optoelectronics and Image Processing, pp. 314–317. Haikou, China (2010)Google Scholar
  19. 19.
    Liu, J., Xiong, Y., Li, B., et al.: Research on automatic detection algorithm of track fastener defects based on computer vision. J. China Railway Soc. 38(8), 73–80 (2016)Google Scholar

Copyright information

© Springer Nature Singapore Pte Ltd. 2019

Authors and Affiliations

  1. 1.Infrastructure Inspection Research Institute, China Academy of Railway Sciences Co., LtdBeijingChina
  2. 2.Beijing IMAP Technology Co., LtdBeijingChina

Personalised recommendations