A Bar Automatic Counting System Based On Dual-View Information Fusion

Bar automatic counting online is a big problem in the iron and steel enterpriseproduction. However, bar counting is still mainly by manual way with the low precision,high cost, low tolerance fault rate, which is beyond the standards desired by enterprise.Therefore, the research on bar counting online system is the key to improve theenterprise production efficiency.Firstly, this paper sum up the four basic problems to note in the time of designingthe automatic counting system embarking from the domestic and foreign research status.Then, the hardware platform of a bar automatic counting system based on dual-viewinformation fusion is designed, which can filter out most of the environmentalinterference.Secondly, in order to take into account the online system fluency and accuracy, theend image processing algorithm are divided into two parts to give respective researchesaccording to the actual end perspective, one is the bar center point detection algorithm inthe case of the high degree stack adhesion, another is the bar center point detectionalgorithm in the case of no stack adhesion.Thirdly, the top image processing algorithm based on Hough transform is proposedto make up the end perspective insufficient information. And the top bar centerrecognition algorithm can accurately and quickly identify bar center under the conditionof no stack adhesion.Lastly, the information fusion rule according to the bar center information form theend and the top is proposed in this paper, and the final bar center point position is got bythe rule. Then by matching the position between the current bar center and the countercenter line we realize the bar counting. And the precision of this system achievedsatisfactory results.