Calibrating a profile measurement system for dimensional inspection in rail rolling mills

被引:5
作者
Millara, Alvaro F. [1 ]
Molleda, Julio [1 ]
Usamentiaga, Ruben [1 ]
Garcia, Daniel F. [1 ]
机构
[1] Univ Oviedo, Dept Comp Sci & Engn, Gijon, Spain
关键词
Camera calibration; Dimensions; Inspection; Machine vision; Optical triangulation; Profile measurement system; Quality control; Rail; Rolling mill;
D O I
10.1007/s00138-020-01147-5
中图分类号
TP18 [人工智能理论];
学科分类号
081104 ; 0812 ; 0835 ; 1405 ;
摘要
Modern, high-speed, railway transportation requires rails to conform strictly to requirements specified in various standards. One key requirement is the conformance of the dimensions of the rail cross section to those of the corresponding rail model, within tight tolerances. This paper deals with a system for dimensional quality inspection during the manufacture of railway rails. Optical triangulation is used to build a profile from laser lines projected on the rails from four different locations. Then, the profile is compared to that of the corresponding rail model. The differences between certain numerical values (the dimensions) for the profile and the model are compared to standard tolerances for each dimension in order to detect dimensional defects. As a prerequisite for this, the cameras used to capture the laser lines must be calibrated. Standard calibration plates are unsuitable for sheet-of-light calibration in a production environment, as determining the location of the laser emitters relative to the plates would be an issue. For this reason, a cylinder-based calibration target is used instead. Different calibration algorithms are discussed and compared to said standard calibration. The results of accuracy and repeatability tests in the production environment are also shown. The accuracy of the system is found to be appropriate for the purpose of quality inspection under the requirements of applicable rail standards.
引用
收藏
页数:16
相关论文
共 22 条
[1]   Fatigue behaviour of rail steel -: a comparison between strain and stress controlled loading [J].
Ahlström, J ;
Karlsson, B .
WEAR, 2005, 258 (7-8) :1187-1193
[2]  
AREMA, 2011, AREMA MAN RAILW ENG
[3]  
European Committee for Standardization, 2011, EN13674
[4]  
Fitzgibbon A. W., 1996, Proceedings of the 13th International Conference on Pattern Recognition, P253, DOI 10.1109/ICPR.1996.546029
[5]  
Gosstandart of Russia, 2001, R516852000 GOST
[6]  
Greenwood Engineering A/S, 2010, MINIPROF DIG PROF ME
[7]  
International Union of Railways, 2008, UIC COD 860 TECHN SP
[8]   Dynamic rail-wear inspecting system based on machine vision [J].
Li, Guoxin ;
Wang, Chonglin ;
Liu, Jianhua ;
Jin, Wenxian .
ICIEA 2007: 2ND IEEE CONFERENCE ON INDUSTRIAL ELECTRONICS AND APPLICATIONS, VOLS 1-4, PROCEEDINGS, 2007, :1-+
[9]   Simple and fast rail wear measurement method based on structured light [J].
Liu, Zhen ;
Sun, Junhua ;
Wang, Heng ;
Zhang, Guangjun .
OPTICS AND LASERS IN ENGINEERING, 2011, 49 (11) :1343-1351
[10]   Absolute three-dimensional shape measurement using coded fringe patterns without phase unwrapping or projector calibration [J].
Lohry, William ;
Chen, Vincent ;
Zhang, Song .
OPTICS EXPRESS, 2014, 22 (02) :1287-1301