基于机器视觉的小方坯端面手写字符自动识别系统  

Automatic handwritten character recognition system for billet end faces based on machine vision

作　　者：石桂芬[1] 何永辉[1] 吴振平[2] SHI Guifen;HE Yonghui;WU Zhenping(Research Institute,Baoshan Iron&Steel Co.,Ltd.,Shanghai 201999,China;Tube,Pipe&Bar Business Unit,Baoshan Iron&Steel Co.,Ltd.,Shanghai 201900,China)

机构地区：[1]宝山钢铁股份有限公司中央研究院,上海201999 [2]宝山钢铁股份有限公司钢管条钢事业部,上海201900

出　　处：《宝钢技术》2022年第3期41-44,共4页Baosteel Technology

摘　　要：针对线材表面质量信息无法按支追溯到坯、对分析异常原因及改进均带来很大局限、造成各种质量异议的问题,结合机组现场条件,自主开发了小方坯端面手写字符自动识别系统。该系统以机器视觉技术为基础,配置高性能双侧照明光源和面扫描CCD图像传感器,获得高质量的方坯端面图像;系统检测识别软件采用快速高效的边缘检测算法和深度学习算法,实时识别出端面手写字符,并将检测结果发送至服务器。系统采用C/S模式为网络架构,实现检测数据在客户端和服务器之间的可靠传递。系统上线1年多的运行实绩证明:该系统在条钢部加热炉前位置,可长期连续工作,按支跟踪准确率达98%以上,实现小方坯的在线按支自动识别和跟踪,具有卓越的检测识别性能和良好的稳定性。During a long time,the surface quality information of wire rod could not be traced back to the billet.The analysis and improvement of abnormal causes great limitations,resulting in a variety of quality objections.Combined with the site conditions of the unit,the automatic handwritten character recognition system was developed to meet the urgent need in Baosteel.This system,which was designed based on machine vision technology,is configured with two illuminations on both sides and an area-array CCD camera as image sensors.High quality images can be obtained on line.Owing to the fast and efficient edge detection algorithm and deep learning algorithm,the images can be processed in real time and the results can be sent to server by the recognition system.The system is based on C/S mode network architecture and the detection data is transmitted between the client and the server.The application performance of the system for more than one year has proved that the system can work continuously for a long time in front of the heating furnace of the bar and steel department,and the tracking accuracy reaches more than 98%.It realizes the online automatic identification and tracking of billet,and has excellent detection and identification performance and good stability.

关 键 词：机器视觉 小方坯 手写字符 自动识别 

分 类 号：TP391.43[自动化与计算机技术—计算机应用技术]