Pattern recognition for infrared profiles of steel strips based on fuzzy knowledge

Author

Usamentiaga, Rubén ; García, Daniel F. ; González, Diego ; Molleda, Julio

Author_Institution

Dept. of Comput. Sci., Oviedo Univ., Asturias

fYear

2006

fDate

12-14 July 2006

Firstpage

122

Lastpage

127

Abstract

The recent demand for extremely thin high-quality steel products makes temperature control an increasingly determining factor in the final quality. In fact, uneven temperature during thin steel production makes the steel fracture rate increase sharply. This work proposes a method to recognize a common uneven temperature pattern known as the hot-shoulders pattern. The proposed recognition method is carried out in three steps. Firstly, the infrared image obtained from the steel strip is processed in order to calculate the infrared profiles. Next, each of these profiles is processed in order to determine its shape. Finally, a fuzzy approach is used to determine the membership degree of each profile to the hot-shoulders temperature pattern

Keywords

feature extraction; fuzzy set theory; image recognition; image segmentation; infrared imaging; steel industry; temperature; feature extraction; fuzzy knowledge; hot-shoulders temperature pattern; image processing; image segmentation; infrared image; infrared profiles; pattern recognition; steel production; steel strips; temperature control; Cooling; Elasticity; Image segmentation; Infrared imaging; Manufacturing; Pattern recognition; Shape; Steel; Strips; Temperature; fuzzy knowledge; image processing; pattern recognition;

fLanguage

English

Publisher

ieee

Conference_Titel

Computational Intelligence for Measurement Systems and Applications, Proceedings of 2006 IEEE International Conference on

Conference_Location

La Coruna

Print_ISBN

1-4244-0244-1

Electronic_ISBN

1-4244-0245-X

Type

conf

DOI

10.1109/CIMSA.2006.250759

Filename

4016841