Research Of Strip Wrinkling Detection And Implementation Of A Monitoring System

Strip wrinkling is a common phenomenon occurring during strip steel annealing procedure. Strip wrinkling is adverse to product quality and cost more material for production. What’s more, when wrinkles hits a severe level the production line may have to shut down and equipment is probably in danger of blemish. Strip wrinkling is one of the biggest obstacles to efficient and steady pipeline for strip steel manufacturer. Research of strip wrinkling and relevant measures has been hot for a long period.To displace the traditional mode of wrinkling detection, which detect strip exception by manpower, this paper introduced two methods to detect strip wrinkling by computer based on digital image processing. A monitoring system which applied the methods has been brought into practice. The contributions of this paper are listed below:1) Introduced a method for online detection of strip wrinkling. This paper investigated the situation of a strip steel manufacturer of Baosteel group, and collect images and data for research. After conducting experiments and explored image processing theories, the author introduced a method for online detection of strip wrinkling. The proposed method consists of four steps:edge detection, strip extraction, image partition and local patch variance computing and wrinkling recognition based on variance matrix. The method proved to be competent for detecting strip wrinkling online precisely.2) To improve the performance of method under varying illumination, this paper introduced a novel approach to detect strip wrinkling. The above method works well under common situations. But is may fail to detect strip wrinkling in extremely high-light condition. The novel method includes four main steps:perspective transformation, strip extraction, key-points detection and wrinkles recognition based on the pattern of key-points located. The novel method has advantage of good performance under high-light condition. Although it takes more computational complexity, it is competent to detect strip wrinkling online.3) A monitoring system is also introduced. The monitoring system applied in a production line adopted the above method to detect strip wrinkling online. The system consist of computer, camera and video acquisition chip。 Most functions are implemented by software. The system is able to detect strip wrinkling and misalignment. Once strip exception is detected, alarm will be triggered to prompt the field worker to react promptly. According to demand of field worker, the system also provides other facilities, such as:video and data preview online, strip situation demonstration, chronological record of video and data, reports for exception. The system improved the condition of strip production and achieved tangible results.