Steel billet recognition in the heavy rail production line

Abstract

Steel billet recognition is an urgent requirement in the steel industry of heavy rail line. Due to high temperature and complex scene in the rolling line, the recognition at the end of billet is quite different from optical character recognition with simple background and vehicle license plate recognition. In order to solve the problem, this paper presents some new schemes in character location, segmentation and recognition in the complex scene. These proposed algorithms include the location algorithm of multilevel filter based on separability criterion, the segmentation algorithm based on intelligent multi-agent and the recognition algorithm of character features matching combining the template matching. The results show that the proposed algorithms are effective for steel billet character recognition in the complex scene. © 2011 Springer-Verlag.