| Strip running deviation is one of the main problems in continuous annealing process line, which would affect the daily production of steel strip severely. Running deviation in the strip continuous production line would invalidate the whole line. It may not only affect productivity but also cause damage to the equipment. Therefore, the study of running deviation is always the hotpot and focus in steel and iron industry.The monitoring system that Baosteel currently uses cannot accurately judge after the positions of camera changed which are caused by the maintenance of devices. This paper firstly introduces a deviation detecting method related to pixel analysis and we find that the ability of recognizing strip deviation has improved. Secondly,the papercomes up with a method which is based on visual attention to detect running deviation.This method cannot only find edge of steel strip automatically, but also resolve the problem of light changing. Finally, we design a comprehensive monitoring system in the paper that includes many functions, such as early warning and reporting of running deviation, to achieve a real-time, pre-alarming and post-completely recording system.This paper introduces functions and realization of image detecting system and presents a better pre-processing with regard to steel running deviation, including cameracalibration, reference setting, bilateral filtering and so forth. Further, the paper comes up with the key techniques of image processing and suggests a method for detecting running deviation to better resolve existing problems.Based on the characteristics of strip steel producing environment, the paper adopts selective visual attention and the algorithms of adaptive thresholdto deal with environment change in continuous production and realize better robustness of running deviation detection. Finally, the paper explains the main functions and real operating effectiveness of the above monitoring system, and suggests the prospect of improvement of the system. |
PDF Full Text Request