摘要
近年来,冶金行业利用智能图像识别技术进行相关应用开发方兴未艾,受冶金行业工艺及装备特点影响,不同的应用点采用的图像识别技术差别很大。介绍一种利用智能图像识别技术的热轧钢卷号自动识别技术,采用图像增强、噪声滤波、目标形态分析、模板匹配、自学习库等技术,在宝钢热轧厂某产线实际应用,取得了良好效果。该技术的应用点在热轧工艺末端的运输链上,是进行卷取运输链区域全自动生产必须解决的技术,可为其他产线提供借鉴。
In recent years,the development of related application technology using intelligent image recognition technology is in the ascendant in chemical and metallurgical industries.But limited by the chemical and metallurgical industries’craft and equipment,the image recognition technologies in different application cases varies greatly.An research in automatic identify the coil number in conventional continuous hot strip mill are introduced.Based on automatic-adaptive gray threshold,image enhancement,image noise filtering and self-learning template library technology,an application is realized and got high recognition accuracy in one hot strip mill in Baosteel.This image recognition technology is used in the transport chain at the end of the hot rolling mill and is one of the key technology to realized the full automatic production.The technical route proposed can be used for reference by other hot strip mills.
作者
赵厚信
陈志荣
ZHAO Houxin;CHEN Zhirong(Hot Rolling Plant,Baoshan Iron&Steel Co.,Ltd.,Shanghai 201900,China)
出处
《宝钢技术》
CAS
2020年第5期41-45,49,共6页
Baosteel Technology
关键词
自适应灰度阀值
图像增强
图像噪声滤波
自学习模板库
adaptive gray threshold
image enhancement
image noise filtering
self-learning library
