Publicaciones Científico Tecnológicas

@conference{Olabarria2008,
title = {Surface Inspection of Hot Laminated Seamless Tubes},
author = {Jose A. Gutiérrez-Olabarria and LAGO MARTINEZ, ALBERTO and CRUZ-LOPEZ CLARET, ANTONIO and AYANI, ALEX},
year = {2008},
date = {2008-02-27},
booktitle = {International Surface Inspection Summit ISIS 2008. Ámsterdam, 27-28/02/2008},
abstract = {This paper describes an application for hot steel tubes surface defects automatic inspection that has been developed for TR SA an important seamless tubes manufacturer company.
At the Hot rolling process when the rolling stands that presses the tubes in order to reduce their diameter are damaged, they sporadically cause surface marks and defects that were not detected at this stage, and this raise the risk of several costly “value added” operations could be performed before detection. Additionally, other metallurgical defects could be present, caused by previous production stages.
The inspection conditions are specially critical due to tube temperature, about 1200ºC, and to tube speed: about 5,5 m/s at the inspection stage. Tube diameters are from 142 mm to 200 mm.
The automated inspection system is installed at the exit of the push bench, just when the defects are originated. This is an especially hard point due to dirt, vapour and temperature conditions, and the system must be capable of inspecting the whole surface of the tubes in real time for defects detection, monitoring and localization of these defects.
The system uses a specially designed illumination, optical filtering and three linear cameras located in three sensorized protective enclosures to inspect the surface, acquiring a complete image of great quality and uniformly illuminated.
An intelligent classification software tool based on Support Vector Machine is used to recognize the previously learned defects.
},
keywords = {computer vision, machine learning, surface quality},
pubstate = {published},
tppubtype = {conference}
}