The Pandrol track fastener image is composed of two parts: track fastener clip sub-graph and track fastener bolt sub-graph. However, the detection of track fastener clip defect can be realized by track fastener image and track fastener image cannot effectively detect whether the bolt is loose. When the convolutional neural network is used to extract whole picture features and detect, many bolt features unrelated to the clips will be obtained, thereby resulting in a high false alarm rate. To solve these problems, a method based on local convolutional neural network to detect the Pandrol track fastener defects is proposed. First, the algorithm for automatic segmentation of track fastener pictures was used to divide the picture of the Pandrol track fastener into two sub-pictures, one sub-picture is the track fastener bolt and the other sub-picture is the track fastener clip. Second, convolutional neural network was used to detect the track fastener clip pictures. The influence of bolt features unrelated to clips on clips detection can be avoided through image segmentation for local feature extraction, thereby reducing the false alarm rate. Finally, the validity of the proposed method is verified using real Pandrol track fastener images.
XiukunWZimingYYuxinL, et al. Railway track fastener defect detection based on image processing and deep learning techniques: a comparative study. Eng Appl Artif Intel2019; 80: 66–81.
2.
GibertXPatelVMChellappaR. Deep multi-task learning for railway track inspection. IEEE Trans Intell Transp2017; 18(1): 153–164.
3.
LvZYanXJiangQ. Batch process monitoring based on self-adaptive subspace support vector data description. Chemometr Intell Lab2017; 170: 25–31.
4.
SongBZhouXShiH, et al. Performance indicator oriented concurrent subspace process monitoring method. IEEE Trans Ind Electron2019; 66: 5535–5545.
5.
KunjieYLiangJQuB, et al. Multiple learning backtracking search algorithm for estimating parameters of photovoltaic models. Appl Energ2018; 226: 408–422.
6.
SongBYanHShiH, et al. Multisubspace elastic network for multimode quality-related process monitoring. IEEE Trans Ind Inform2020; 16: 5874–5883.
7.
LvZYanXJiangQ, et al. Just-in-time learning-multiple subspace support vector data description used for non-Gaussian dynamic batch process monitoring. J Chemometr2019; 33(6): e3134.
8.
FengHJiangZXieF, et al. Automatic fastener classification and defect detection in vision-based railway inspection systems. IEEE Trans Instrum Meas2014; 63(4): 877–888.
9.
PengTTaoHYangC, et al. A uniform modeling method based on open-circuit faults analysis for NPC-three-level converter. IEEE Trans Circuits Syst II2019; 66(3): 457–461.
10.
MhaskarHNPoggioT.Deep vs. shallow networks: an approximation theory perspective. Anal Appl2016; 14(6): 829–848.
11.
LecunYBengioYHintonG. Deep learning. Nature2015; 521(7553): 436–444.
12.
WuZTKarimiHRDangCY. A deterministic annealing neural network algorithm for the minimum concave cost transportation problem. IEEE Trans Neur Net Lear. Epub ahead of print 19 December 2019. DOI: 10.1109/TNNLS.2019.2955137.
13.
FanHCosmanPCHouY, et al. High-speed railway fastener detection based on a line local binary pattern. IEEE Signal Proc Let2018; 25: 788–792.
14.
NaielMAAhmadMOSwamyMN. A vehicle detection scheme based on two-dimensional HOG features in the DFT and DCT domains. Multidim Syst Sign Process2019; 30: 1697–1729.
15.
DalalNTriggsB. Histograms of oriented gradients for human detection. In: 2005 IEEE computer society conference on computer vision and pattern recognition (CVPR’05), San Diego, CA, 20–25 June 2005. New York: IEEE.
16.
PavaniSKDelgado-GomezDFrangiAF. Gaussian weak classifiers based on co-occurring Haar-like features for face detection. Pattern Anal Appl2014; 17(2): 431–439.
17.
XiaYXieFJiangZ. Broken railway fastener detection based on AdaBoost algorithm. In: International conference on optoelectronics & image processing, Haikou, China, 11–12 November 2011. New York: IEEE.
18.
Dollar Piotr ZitnickCL. Fast edge detection using structured forests. IEEE Trans Pattern Anal2014; 37(8): 1558–1570.
19.
WangZLiuHWangZ, et al. Scale-invariant feature matching based on pairs of feature points. IET Comput Vis2015; 9(6): 789–796.
20.
ChenXLiuLDengY, et al. Vehicle detection based on visual attention mechanism and AdaBoost cascade classifier in intelligent transportation systems. Opt Quant Electron2019; 51(8): 263.
21.
TrinhHHaasNLiY, et al. Enhanced rail component detection and consolidation for rail track inspection. In: 2012 IEEE workshop on the applications of computer vision, Breckenridge, CO, 9–11 January 2012. New York: IEEE.
22.
GouJPMaHX. A generalized mean distance-based k-nearest neighbor classifier. Expert Syst Appl2019; 115: 356–372.
23.
TyagiKManryM. Multi-step training of a generalized linear classifier. Neural Process Lett2019; 50: 1341–1360.
24.
LiuLZhouFQHeYZ. Automated status inspection of fastening bolts on freight trains using a machine vision approach. Proc IMechE, Part F: J Rail Rapid Transit2016; 230(7): 1629–1641.
25.
ResendizEHartJAhujaN. Automated visual inspection of railroad tracks. IEEE Trans Intell Transp2013; 14(2): 751–760.
26.
GibenXPatelVMChellappaR. Material classification and semantic segmentation of railway track images with deep convolutional neural networks. In: IEEE international conference on image processing, Quebec City, QC, Canada, 27–30 September 2015. New York: IEEE.
27.
KunjieYBoyangQCaitongY, et al. A performance-guided JAYA algorithm for parameters identification of photovoltaic cell and module. Appl Energ2019; 237: 241–257.
28.
KunjieYLiangJQuB, et al. Parameters identification of photovoltaic models using an improved JAYA optimization algorithm. Energ Convers Manage2017; 150: 742–753.
29.
LiYTrinhHHaasN, et al. Rail component detection, optimization, and assessment for automatic rail track inspection. IEEE Trans Intell Transp2014; 15(2): 760–770.
30.
WuZTKarimiHRDangCY. An approximation algorithm for graph partitioning via deterministic annealing neural network. Neural Netw2019; 117: 191–200.
31.
HeKZhangXRenS, et al. Deep residual learning for image recognition. In: 2016 IEEE conference on computer vision and pattern recognition (CVPR), Las Vegas, NV, 27–30 June 2016. New York: IEEE.
32.
SzegedyCLiuWJiaY, et al. Going deeper with convolutions. In: 2015 IEEE conference on computer vision and pattern recognition (CVPR), Boston, MA, 7–12 June 2015. New York: IEEE.
33.
ChenJLiuZWangH, et al. Automatic defect detection of fasteners on the catenary support device using deep convolutional neural network. IEEE Trans Instrum Meas2018; 67(2): 257–269.
34.
LiuJBHuangYPZouQ, et al. Learning visual similarity for inspecting defective railway fasteners. IEEE Sens J2019; 19(16): 6844–6857.
35.
ChenYLinZZhaoX, et al. Deep learning-based classification of hyperspectral data. IEEE J Stars2014; 7(6): 2094–2107.
36.
LiangHSunXSunY, et al. Text feature extraction based on deep learning: a review. EURASIP J Wirel Comm2017; 2017(1): 211.
37.
MirjaliliSAHashimSZMSardroudiHM. Training feedforward neural networks using hybrid particle swarm optimization and gravitational search algorithm. Appl Math Comput2012; 218(22): 11125–11137.
38.
WeiHYangyuHLiW, et al. Deep convolutional neural networks for hyperspectral image classification. J Sens2015; 2015: 258619.
39.
KananCCottrellGW. Color-to-grayscale: does the method matter in image recognition. PLoS ONE2012; 7(1): 133–139.
PanB. Bias error reduction of digital image correlation using Gaussian pre-filtering. Opt Laser Eng2013; 51(10): 1161–1167.
42.
Debella-GiloMKaabA. Sub-pixel precision image matching for measuring surface displacements on mass movements using normalized cross-correlation. Remote Sens Environ2011; 115(1): 130–142.
43.
YangCGuiWChenZ, et al. Voltage difference residual-based open-circuit fault diagnosis approach for three-level converters in electric traction systems. IEEE Trans Power Electr2020; 35(3): 3012–3028.
44.
ZhiwenCXuemingLChaoY, et al. A data-driven ground fault detection and isolation method for main circuit in railway electrical traction system. ISA Trans2019; 87: 264–271.
45.
DuanLTsangIWXuD. Domain transfer multiple Kernel learning. IEEE Trans Pattern Anal2012; 34(3): 465–479.
46.
XueJHHallP. Why does rebalancing class-unbalanced data improve AUC for linear discriminant analysis?IEEE Trans Pattern Anal2015; 37(5): 1109–1112.
