The Research And Application Of Separating Steel Bars Automatically In Rolling Mills Based On Machine Vision

The main production of the rolling mills, or the bar mills, is steel and round steel. The rolling mills generally use the sale of the actual weight of a bundle of steel bars. Either way of the sale styles, there are two essential works. One is to count the number of a bundle of steel bars; the other is to split the steel bars between two bundles of steel bars. Most of rolling mills use artificially manual counting and separation manner, this approach is prone to statistical errors and separation errors. On the one hand it will bring a direct impact on economic to rolling mills, on the other hand it will reduce the credibility of rolling mills if the situation appears.This paper presents an efficient and easy way for separating steel bars automatically, which is called the system of separating steel bars automatically. It adopts advanced image recognition technology to identify the end face of bars and count the number of the bars. When the statistics of the number of bars reaches the rating number, the chain bed is stopped by IPC's command and the IPC sends separation of steel bars signals to servo driver through serial port communication. The servo driver controls the servo motor driving separation machine to the appointed separation position and separates steel bars. After that, the IPC controls the middle separation machine to widen the distance between the steel bars until the separation is completed. When the bars are totally separated, the separation machine send a signal to IPC, which will start the next automatic count and separation process. This is the whole separation process.Automatic separating steel bars system is implemented on the basis of the automatic counting system. The automatic separating steel bars system has been debugged successfully in NO.1 workshop of Laiwu rolling mill in Shandong province and has been used formerly. The automatic counting system adopts collecting images of bar end to identify the bar information. In this paper, the cheer algorithm, which is widely used in image processing, is used to recognition the bar center. The cheer algorithm is circular-like target recognition algorithm, which uses edge detection to obtain the edge information of the end face of steel bars and direction information. Then it enhances the central message of the end face of steel bars and uses central aggregation algorithm to get and confirm the center of the end face of steel bars next. At last, it establishes circular-like center according to certain parameters.The research of the automatic separating steel bars system has two main innovations:1) This paper uses a high-precision and fast servo motor to achieve the precise and rapid separation of steel bars, which guarantees the correctness of the number of bundles of bar and do not delay the normal production rhythm of the factory workshop in rolling mills. The accuracy of the servo motor can achieve micron-level, and its speed can reach 3000r/min.2) This paper proposes a design method of separation emendation to resolve the complex situation of the multiple stacked steel bars. Separation emendation is adopted firstly to make the distance of the bars become large in order to separate them simply. Separation emendation takes the measure of controlling the end separation wedge raise a distance aiming at pulling the bars away. This is undoubtedly the bottleneck of the automatic separating steel bars'system. If this problem is resolved, the separation system will achieve a very high level.