冷轧钢表面与内部缺陷检测研究  

Detection of surface and internal defects in cold rolled steel

作　　者：陈名渝 谢玥辰 吕雄涛 郭建荣 贾国军[4] 许志鹏 王狮凌 项震[1] 刘东[1,2] CHEN Ming-yu;XIE Yue-chen;LV Xiong-tao;GUO Jian-rong;JIA Guo-jun;XU Zhi-peng;WANG Shi-ling;XIANG Zhen;LIU Dong(State Key Laboratory of Extreme Photonics and Instrumentation,College of Optical Science and Engineering,Zhejiang University,Hangzhou 310027,China;ZJU-Hangzhou Global Scientific and Technological Innovation Center,Hangzhou 311200,China;Zhejiang Southeast New Material Technology Company Limited,Hangzhou 311222,China;Technology Center,Ningbo Steel Company Limited,Ningbo 315807,China;College of Metrology and Measurement Engineering,China Jiliang University,Hangzhou 310018,China)

机构地区：[1]浙江大学光电科学与工程学院极端光学技术与仪器全国重点实验室,浙江杭州310027 [2]浙江大学杭州国际科创中心,浙江杭州311200 [3]浙江东南新材科技有限公司,浙江杭州311222 [4]宁波钢铁有限公司技术中心,浙江宁波315807 [5]中国计量大学计量测试工程学院,浙江杭州310018

出　　处：《中国光学（中英文）》2024年第4期823-833,共11页Chinese Optics

基　　金：国家重点研发计划(No.2022YFB3403404);极端光学技术与仪器全国重点实验室创新项目(No.EPI2023ZD01)。

摘　　要：为实现冷轧钢缺陷的全面检测,针对其表面和内部缺陷检测展开研究。对于表面缺陷检测,提出采用双侧线光源照明方案,并与常规线光源照明方案进行对比。对于内部缺陷检测,从检测分辨率和缺陷边缘特征两方面分析X射线、超声以及红外热波成像等金属内部检测技术的适用性。经实验验证,双侧线光源照明不仅可以使YOLOv5目标检测算法总体平均精度mAP:0.5达到90.16%,相比线光源照明提升了15.46%,还可优化模型分类和提高训练效率。X射线和超声波检测法可检测直径为0.25 mm的盲孔,而红外热波成像技术则可有效识别出直径为1 mm的盲孔。在缺陷边缘特征评估中,X射线检测法的最小盲孔边缘灰度差值为145,超声波为89,红外热波成像为30。本研究提出了一种冷轧钢表面缺陷检测的改进方案,并为内部缺陷检测提供了思路。This paper focuses on the comprehensive detection of defects in cold rolled steel through examina-tion for surface and internal defects.Regarding surface defect detection,a bilateral line light illumination scheme is proposed and a comparison with line light illumination scheme is carried out.As for internal de-fect detection,the applicability of various metal internal inspection technologies such as X-ray,ultrasound,and infrared thermography is analyzed from the perspectives of detection resolution and defect edge charac-teristics.The results show that bilateral line light illumination not only increases the overall average preci-sion of the YOLOv5 object detection algorithm model to 90.16%(an increase of 15.46%compared to the line light illumination)but also improves model classification and training efficiency.X-ray and ultrasound inspection technologies can detect blind holes with a diameter of 0.25 mm,while infrared thermography can detect blind holes with a diameter of 1 mm.In evaluating defect edge characteristics,X-ray inspection tech-nology exhibits a minimum blind hole edge grayscale difference of 145,ultrasound of 89,and infrared ther-mography of 30.This study proposes an improved scheme for the detection of surface defects in cold rolled steel and offers insights for the research on internal defect detection.

关 键 词：冷轧钢缺陷检测 表面缺陷 内部缺陷 双侧线光源照明 

分 类 号：O439[机械工程—光学工程]