DeepRail: Automatic Visual Detection System for Railway Surface Defect Using Bayesian CNN and Attention Network

被引：0
|
作者：
Jin X.-T. [1 ,2 ]
Wang Y.-N. [1 ,2 ]
Zhang H. [2 ,3 ]
Liu L. [1 ,2 ]
Zhong H. [1 ,2 ]
He Z.-D. [4 ]
机构：
[1] College of Electrical and Information Engineering, Hunan University, Changsha
[2] National Engineering Laboratory of Robot Vision Perception and Control Technology, Hunan University, Changsha
[3] College of Electrical and Information Engineering, Changsha University of Science and Technology, Changsha
[4] College of Electrical and Information Engineering, Zhengzhou University of Light Industry, Zhengzhou
来源：
Zidonghua Xuebao/Acta Automatica Sinica | 2019年 / 45卷 / 12期
基金：
中国国家自然科学基金;
关键词：
Attention mechanism; Bayesian convolutional neural network (CNN); Class imbalance; Rail surface defects; Visual detection;
D O I：
10.16383/j.aas.c190143
中图分类号：
学科分类号：
摘要：
This paper extends the state-of-the-art deep learning framework DeepLab v3+ to a light-weighted and scalable Bayesian version DeeperLab for the defect detection on complex and diverse rail surface. Specifically, Dropout is incorporated into the improved Xception network for Monte Carlo sampling from posterior distribution. Atrous spatial pyramid pooling (ASPP) module is utilized to extract the dense features at multiple scales and rates. Furthermore, a simpler and efficient decoder is proposed to improve the defect edges, and outputs the mean and variance of Softmax probability as segmentation and uncertainty. To solve class imbalance problem, we present the loss attention network (LAN) to perform auxiliary supervision for DeeperLab training. Experimental results show that the proposed algorithm is more accurate and robust than other methods with 91.46 % precision and 0.18 s/frame speed. Copyright © 2019 Acta Automatica Sinica. All rights reserved.
引用
收藏
页码：2312 / 2327
页数：15
相关论文
共 36 条
  • [1] He Z.-D., Wang Y.-N., Mao J.-X., Yin F., Research on inverse P-M diffusion-based rail surface defect detection, Acta Automatica Sinica, 40, 8, pp. 1667-1679, (2014)
  • [2] He Z.D., Wang Y.N., Yin F., Liu J., Surface defect detection for high-speed rails using an inverse PM diffusion model, Sensor Review, 36, 1, pp. 86-97, (2016)
  • [3] Resendiz E., Hart J.M., Ahuja N., Automated visual inspection of railroad tracks, IEEE Transactions on Intelligent Transportation Systems, 14, 2, pp. 751-760, (2013)
  • [4] Sun C.-S., Zhang Y.-H., Research on automatic early warning method for rail flaw based on intelligent identification and periodic detection, Journal of the China Railway Society, 40, 11, pp. 140-146, (2018)
  • [5] Liang B., Iwnicki S., Ball A., Young A.E., Adaptive noise cancelling and time-frequency techniques for rail surface defect detection, Mechanical Systems and Signal Processing, 54-55, pp. 41-51, (2015)
  • [6] Gibert X., Patel V.M., Chellappa R., Deep multitask learning for railway track inspection, IEEE Transactions on Intelligent transportation systems, 18, 1, pp. 153-164, (2017)
  • [7] Giben X., Patel V.M., Chellappa R., Material classification and semantic segmentation of railway track images with deep convolutional neural networks, Proceedings of the 2015 IEEE International Conference on Image Processing (ICIP), pp. 621-625, (2015)
  • [8] Faghih-Roohi S., Hajizadeh S., Nunez A., Babuska R., Deep convolutional neural networks for detection of rail surface defects, Proceedings of the 2016 International Joint Conference on Neural Networks (IJCNN), pp. 2584-2589, (2016)
  • [9] Masci J., Meier U., Ciresan D., Et al., Steel defect classification with max-pooling convolutional neural networks, Proceedings of the 2012 International Joint Conference on Neural Networks (IJCNN), pp. 1-6, (2012)
  • [10] Chen J.W., Liu Z.Y., Wang H.R., Nunez A., Han Z.W., Automatic defect detection of fasteners on the catenary support device using deep convolutional neural network, IEEE Transactions on Instrumentation and Measurement, 67, 2, pp. 257-269, (2018)
1234