Algorithm Research For Steel Billet Characters Positioning On The Production Line

Characters recognition is a new automation technology. It is always used to identify some special symbols such as characters, letters, numbers, patterns, etc., and it can intelligently extract the recognition results. Because of its unique intelligence, it has been widely used in many fields, such as smart parking license plate recognition, fingerprint recognition, face recognition, characters recognition. The main research of this paper is to identify the steel billet characters. Characters recognition generally include pretreatment, characters positioning, characters segmentation, and characters recognition, this paper aims to study the algorithm of billet characters positioning. Characters positioning accuracy is directly related to the follow-up work, characters positioning is a very important step of characters recognition.This article is originated from the project of heavy rail steel production in Wuhan Iron and Steel(Group) Company. Billet characters record the information of the billets, using computer technology and image processing method to recognize and improve the production efficiency, is the significance of this paper. Billet ends has its own unique characters. The printing characters on billet are white, yet the billet itself is black or gray, and the the environment of workshop is mostly gray. So in the collected images, backgrounds and characters are difficult to separate, and the processing of image backgrounds is a difficult work, we need to find an appropriate method for character positioning.To solve the above problems, firstly, this paper studies the differences between billet characters positioning and the traditional characters positioning. For the pictures collected on the steel billet production line, due to various factors such as light, noise and other interference, the background and characters area are difficult to separate. So this paper proposes a reasonable solution with Mean shift filtering method and OTSU segmentation method, characters positioning can be realized. This paper focus on the billet image segmentation and filtering method.Secondly, an algorithm of slant correction for billet characters is also proposed.Aiming at the problem that the slant angle of the billet image is large, this algorithm get good results.Finally, the algorithm for billet characters positioning is summarized and analyzed, the problems of the algorithm and possible solutions is proposed.On the base of a large number of experimental results, it shows that the algorithm accurately located the billet characters, it successfully resolved many problems of the billet characters recognition on the production line, and it provides a basis for billet characters recognition.