Research of method for detection of rail fastener defects based on machine vision

DOI

10.2991/icmmcce-15.2015.547How to use a DOI?

Keywords

railway fasteners, Canny operator, Hough transform, fusion of LBP and HOG,SVM.

Abstract

The detection of rail fastener defects is the key to ensure the running safety of high-speed trains. Traditional method is usually be detected rely on train workers who walk along railway lines to find out the potential risks. The method by artificial maintenance is slowly, costly, and dangerous. As to solve the problem, an automatic detect method based on machine vision is proposed for all kinds of rail fastener defects. background subtraction algorithm is used to achieve the accurate positioning of railway fasteners in this paper. The method is based on video sequence for processing rail fastener image. First of all, realized rail fastener accurate positioning based on the improved Canny operator and the Hough transform to extract linear; Then, extracted the characteristics of fastener defect feature by fusion LBP (Local Binary Pattern, Local Binary Pattern) and HOG (Histogram of Oriented Gradient direction Gradient Histogram); Finally, the characteristics of Histogram as the SVM (Support Vector Machine SVM) input values is used to buckle defect classification. The experimental results show that the method compared with the traditional single classification method, has higher real-time and accuracy, and can meet the requirement of rail fastener defect detection.

Copyright

© 2015, the Authors. Published by Atlantis Press.

Open Access

This is an open access article distributed under the CC BY-NC license (http://creativecommons.org/licenses/by-nc/4.0/).

Download article (PDF)

TY  - CONF
AU  - Zhenzhen Wang
AU  - Siming Wang
PY  - 2015/12
DA  - 2015/12
TI  - Research of method for detection of rail fastener defects based on machine vision
BT  - Proceedings of the 4th International Conference on Mechatronics, Materials, Chemistry and Computer Engineering 2015
PB  - Atlantis Press
SN  - 2352-538X
UR  - https://doi.org/10.2991/icmmcce-15.2015.547
DO  - 10.2991/icmmcce-15.2015.547
ID  - Wang2015/12
ER  -