冷轧钢表面与内部缺陷检测研究

陈名渝; 谢玥辰; 吕雄涛; 郭建荣; 贾国军; 许志鹏; 王狮凌; 项震; 刘东

doi:10.37188/CO.2023-0189

冷轧钢表面与内部缺陷检测研究

doi: 10.37188/CO.2023-0189

陈名渝^{1, 2,},
谢玥辰^{1, 2},
吕雄涛^{1, 2},
郭建荣³,
贾国军⁴,
许志鹏²,
王狮凌⁵,
项震¹,
刘东^{1, 2, ,}

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

基金项目: 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）

详细信息

作者简介:
刘　东（1982—），男，辽宁大连人，博士，教授，博士生导师，浙江大学光电科学与工程学院副院长，极端光学技术与仪器全国重点实验室副主任，中国光学工程学会理事，中国光学学会激光光谱学专委会副主任委员，《大气与环境光学学报》执行副主编、《激光技术》编委会副主任委员、《光学学报》等期刊编委。担任多个国际/国内学术会议主席/共主席等。主要研究方向包括机器视觉缺陷检测、极端应用干涉检测、环境探测激光雷达等。主持国家重点研发计划项目及国家自然科学基金项目4项，出版教材及专著5部，作为第一/通讯作者在PNAS、Research、PhotoniX、Light: Sci & App等期刊上发表学术论文百余篇，授权的国家发明专利实现成果转化10余项, 国内外学术会议作大会/主旨/邀请报告70余次。研究成果获浙江省科技进步一等奖1项（排名第一）、“金穗奖”中国光电仪器品牌榜金奖1项（排名第一）等。E-mail：liudongopt@zju.edu.cn

中图分类号: O439
计量
- 文章访问数: 85
- HTML全文浏览量: 56
- PDF下载量: 18
- 被引次数: 0
出版历程
- 收稿日期: 2023-10-31
- 修回日期: 2023-11-24
- 网络出版日期: 2024-05-09

Detection of surface and internal defects in cold rolled steel

CHEN Ming-yu^{1, 2
,},
XIE Yue-chen^{1, 2},
LV Xiong-tao^{1, 2},
GUO Jian-rong³,
JIA Guo-jun⁴,
XU Zhi-peng²,
WANG Shi-ling⁵,
XIANG Zhen¹,
LIU Dong^{1, 2
, ,}

1.
State Key Laboratory of Extreme Photonics and Instrumentation, College of Optical Science and Engineering, Zhejiang University, Hangzhou 310027, China
2.
ZJU-Hangzhou Global Scientific and Technological Innovation Center, Hangzhou 311200, China
3.
Zhejiang Southeast New Material Technology Company Limited, Hangzhou 311222, China
4.
Technology Center, Ningbo Steel Company Limited, Ningbo 315807, China
5.
College of Metrology and Measurement Engineering, China Jiliang University, Hangzhou 310018, China

Funds: Supported by National key research and development program (No. 2022YFB3403404); State Key Laboratory of Extreme Photonics and Instrumentation Innovation Program (No. EPI2023ZD01)

More Information

Corresponding author: liudongopt@zju.edu.cn

摘要

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。
- 冷轧钢缺陷检测 /
- 表面缺陷 /
- 内部缺陷 /
- 双侧线光源照明
Abstract:
This paper focuses on the comprehensive detection of defects in cold rolled steel through examination 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 defect 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 characteristics. The results show that bilateral line light illumination not only increases the overall average precision 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 technology exhibits a minimum blind hole edge grayscale difference of 145, ultrasound of 89, and infrared thermography 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.
- cold rolled steel defect detection /
- surface defects /
- internal defects /
- bilateral line light illumination

HTML全文

图 1 冷轧钢缺陷检测方案

Figure 1. Cold rolled steel defect detection scheme

下载: 全尺寸图片幻灯片

图 2 冷轧钢表面检测模型

Figure 2. Model of cold rolled steel surface detection

下载: 全尺寸图片幻灯片

图 3 不同相机拍摄角${\theta _r}$下的检测划痕

Figure 3. Detecting scratches captured at different angles ${\theta _r}$ by camera

下载: 全尺寸图片幻灯片

图 4 3种线光源照明方向示意图

Figure 4. Lighting direction of three kinds of line light sources

下载: 全尺寸图片幻灯片

图 5 不同光源下各方向划痕仿真结果

Figure 5. Simulation results of scratches illuminated by different light sources

下载: 全尺寸图片幻灯片

图 6 冷轧钢表面缺陷检测系统

Figure 6. The cold rolled steel surface defect detection system

下载: 全尺寸图片幻灯片

图 7 盲孔加工样品

Figure 7. Sample with blind hole

下载: 全尺寸图片幻灯片

图 8 样品缺陷识别结果

Figure 8. Defect identification results

下载: 全尺寸图片幻灯片

图 9 针对不同缺陷时，白光线光源与白光双侧线光源照明下Precision-Recall曲线。（a）夹杂；（b）树纹；（c）划痕

Figure 9. Precision-Recall curve for different defects under white line light source and white bilateral line light source illumination. (a) Inclusion; (b) tree grain; (c) scratch

下载: 全尺寸图片幻灯片

图 10 白光线光源与白光双侧线光源照明下平均精度mAP:0.5随训练轮数变化关系

Figure 10. Variation of mAP:0.5 with the epochs under white line light source and white bilateral line light source illumination

下载: 全尺寸图片幻灯片

图 11 内部缺陷检测结果

Figure 11. Detection results of internal defects

下载: 全尺寸图片幻灯片

图 12 X射线检测几何放大原理

Figure 12. Geometric magnification principle for X-ray detection

下载: 全尺寸图片幻灯片

图 13 Sobel算法盲孔样品边缘特征提取

Figure 13. Blind hole sample edge detection using Sobel algorithm

下载: 全尺寸图片幻灯片

表 1 冷轧钢样品缺陷尺寸

Table 1. Defect size in cold rolled steel samples

缺陷类型	长度均值/cm	不确定度	宽度均值/cm	不确定度
孔洞	0.896	0.035	0.512	0.020
破损	4.960	0.025	1.204	0.030
夹杂	15.52	0.74	0.162	0.091
树纹	8.242	0.030	0.200	0.025
划痕	20.49	0.74	0.0222	0.0084
色差	19.70	0.74	5.994	0.030

下载: 导出CSV

表 2 白光线光源与白光双侧线光源之间的对比

Table 2. Imaging comparison when using white lateral and white bilateral line light source illuminations

	孔洞	破损	夹杂	树纹	纵向划痕	横向划痕	色差
白光单侧线光源
白光双侧线光源

下载: 导出CSV

表 3 白光线光源与白光双侧线光源照明下YOLOv5目标检测算法结果

Table 3. Results of YOLOv5 target detection algorithm using white line light source and white bilateral line light source illuminations

	白光线光源	白光双侧线光源
准确率	80.80%	91.50%↑
召回率	96.00%	97.67%↑
mAP:0.5	74.70%	90.16%↑
损失值	1.38%	1.37%
孔洞	99.50%	99.50%
破损	99.60%	99.60%
夹杂	74.80%	89.80%↑
树纹	74.60%	96.27%↑
划痕	65.57%	83.90%↑
色差	\	73.00%

下载: 导出CSV

表 4 Sobel算法盲孔样品边缘特征提取结果

Table 4. Results of edge detection on blind hole sample by applying the Sobel algorithm

	工业CT	超声检测	红外热波成像
盲孔检出数	20	20	15
盲孔边缘连通数	20	19	13
边缘提取准确数	20	19	10
最大盲孔边缘灰度差值
φ2.50 mm	190	193	104
φ1.75 mm	187	180	80
φ1.00 mm	184	172	30
φ0.25 mm	145	89	/

下载: 导出CSV

参考文献(30)

[1]	LI ZH, LI ZH, HUANG H, et al. Green lithium: photoelectrochemical extraction[J]. PhotoniX, 2023, 4(1): 23.
[2]	ZHU Y N, LUO H, YANG CH Y, et al. Color-preserving passive radiative cooling for an actively temperature-regulated enclosure[J]. Light: Science & Applications, 2022, 11(1): 122.
[3]	李玮. 用于锂离子电池钢壳的镀镍钴钢带制备工艺及耐蚀性研究[D]. 湘潭: 湘潭大学, 2011. LI W. The study of preparation and corrosion resistance of cobalt-nickel-plated steel strip for lithium-ion battery shell[D]. Xiangtan: Xiangtan University, 2011. (in Chinese).
[4]	赵小龙, 王雄, 罗晓阳, 等. 退火温度对CSP冷轧低碳高强钢带组织和性能的影响[J]. 金属热处理,2021,46(1):43-49. ZHAO X L, WANG X, LUO X Y, et al. Effect of annealing temperature on microstructure and properties of cold rolled low carbon high strength sheet by CSP[J]. Heat Treatment of Metals, 2021, 46(1): 43-49. (in Chinese).
[5]	MANCINI E, CAMPANA F, SASSO M, et al. Effects of cold rolling process variables on final surface quality of stainless steel thin strip[J]. The International Journal of Advanced Manufacturing Technology, 2012, 61(1-4): 63-72.
[6]	董双鹏, 宋进英, 陈业熊, 等. 冷轧钢板表面缺陷的种类及形成原因[J]. 机械工程材料,2012,36(7):100-104. DONG SH P, SONG J Y, CHEN Y X, et al. Types and forming reasons of cold rolled steel plate surface defects[J]. Materials for Mechanical Engineering, 2012, 36(7): 100-104. (in Chinese).
[7]	胡华东, 王乐, 汤化胜, 等. 冷轧宽厚规格深冲钢梗印缺陷成因分析及其控制措施[J]. 轧钢,2022,39(2):124-128. HU H D, WANG L, TANG H SH, et al. Cause analysis of stem seal defects of cold rolled wide and thick gauge deep drawing steel strip and its control measures[J]. Steel Rolling, 2022, 39(2): 124-128. (in Chinese).
[8]	王畅, 于洋, 王林, 等. 冷轧双相钢镀锌条带状色差缺陷产生原因及机理研究[J]. 轧钢,2023,40(1):52-58. WANG CH, YU Y, WANG L, et al. Study on the cause and mechanism of the strip color difference defects of cold rolled dual- phase steel galvanized strip[J]. Steel Rolling, 2023, 40(1): 52-58. (in Chinese).
[9]	SURESH B R, FUNDAKOWSKI R A, LEVITT T S, et al. A real-time automated visual inspection system for hot steel slabs[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 1983, PAMI-5(6): 563-572.
[10]	NATH V, CHATTOPADHYAY C, DESAI K A. On enhancing prediction abilities of vision-based metallic surface defect classification through adversarial training[J]. Engineering Applications of Artificial Intelligence, 2023, 117: 105553.
[11]	杨水山, 何永辉, 王康健. 高速冷连轧带钢表面质量在线检测系统开发与应用[J]. 世界钢铁,2013,13(3):62-67. doi: 10.3969/j.issn.1672-9587.2013.03.011 YANG SH SH, HE Y H, WANG K J. Development on on-line inspection system for high speed strip surface quality at cold continuous rolling mill[J]. World Iron & Steel, 2013, 13(3): 62-67. (in Chinese). doi: 10.3969/j.issn.1672-9587.2013.03.011
[12]	沈春光, 李虎威, 荆涛, 等. 基于深度学习的带钢表面缺陷检测在小样本数据集的应用[J]. 轧钢,2022,39(2):82-86. SHEN CH G, LI H W, JING T, et al. Application of deep learning-based surface defect recognition of plate in small sample dataset[J]. Steel Rolling, 2022, 39(2): 82-86. (in Chinese).
[13]	JUNet A, MESSAGER A, WECK A, et al. Internal fatigue crack propagation in a Ti-6Al-4V alloy: an in situ study[J]. International Journal of Fatigue, 2023, 168: 107450.
[14]	YANG Q ZH, ZHAO Y, HU X A, et al. Small crack propagation and closure behaviours and mechanisms in a powder metallurgy superalloy at high temperature in air[J]. International Journal of Fatigue, 2023, 175: 107797.
[15]	白贇沨, 刘祎, 张小琳, 等. 基于GF-WLS和VSM的多能量X射线图像融合方法[J]. 无损检测,2022,44(9):34-41,74. doi: 10.11973/wsjc202209007 BAI Y F, LIU Y, ZHANG X L, et al. Multi energy X-ray image fusion method based on GF-WLS and VSM[J]. Nondestructive Testing, 2022, 44(9): 34-41,74. (in Chinese). doi: 10.11973/wsjc202209007
[16]	TAKADA H, TOMURA Y, ARATANI M, et al. On-line detection system for internal flaws in as-hot-rolled steel strip using ultrasonic probe array[J]. Materials Transactions, 2011, 52(3): 531-538.
[17]	张永生, 黄松岭, 赵伟, 等. 基于电磁超声的钢板裂纹检测系统[J]. 无损检测,2009,31(4):307-310. ZHANG Y SH, HUANG S L, ZHAO W, et al. Detection system about steel cracks based on electronic ultrasonic[J]. Nondestructive Testing, 2009, 31(4): 307-310. (in Chinese).
[18]	LIU SH Y, JIA K N, WAN H P, et al. Inspection of the internal defects with different size in Ni and Ti additive manufactured components using laser ultrasonic technology[J]. Optics & Laser Technology, 2022, 146: 107543.
[19]	ALMOND D P, PICKERING S G. An analytical study of the pulsed thermography defect detection limit[J]. Journal of Applied Physics, 2012, 111(9): 093510.
[20]	ZHUO L J, YANG X, ZHU J G, et al. Size determination of interior defects by reconstruction of subsurface virtual heat flux for step heating thermography[J]. NDT & E International, 2023, 133: 102734.
[21]	WANG F, SHENG J, SFARRA S, et al. Multimode infrared thermal-wave imaging in non-destructive testing and evaluation (NDT&E): physical principles, modulated waveform, and excitation heat source[J]. Infrared Physics & Technology, 2023, 135: 104993.
[22]	NAYAR S K, IKEUCHI K, KANADE T. Surface reflection: physical and geometrical perspectives[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 1991, 13(7): 611-634.
[23]	张青树, 齐韦. 冷轧带钢表面形貌测试及其粗糙度均匀性控制研究[J]. 轧钢,2022,39(5):103-107. ZHANG Q SH, QI W. Study on surface profile test and roughness uniformity control of cold rolled strip[J]. Steel Rolling, 2022, 39(5): 103-107. (in Chinese).
[24]	杨甬英, 陆春华, 梁蛟, 等. 光学元件表面缺陷的显微散射暗场成像及数字化评价系统[J]. 光学学报,2007,27(6):1031-1038. doi: 10.3321/j.issn:0253-2239.2007.06.015 YANG Y Y, LU CH H, LIANG J, et al. Microscopic dark-field scattering imaging and digitalization evaluation system of defects on optical devices precision surface[J]. Acta Optica Sinica, 2007, 27(6): 1031-1038. (in Chinese). doi: 10.3321/j.issn:0253-2239.2007.06.015
[25]	LI X C, WU Q Q, XIAO B P, et al. High-speed and robust infrared-guiding multiuser eye localization system for autostereoscopic display[J]. Applied Optics, 2020, 59(14): 4199-4208.
[26]	吴海滨, 魏喜盈, 刘美红, 等. 结合空洞卷积和迁移学习改进YOLOv4的X光安检危险品检测[J]. 中国光学,2021,14(6):1417-1425. WU H B, WEI X Y, LIU M H, et al. Improved YOLOv4 for dangerous goods detection in X-ray inspection combined with atrous convolution and transfer learning[J]. Chinese Optics, 2021, 14(6): 1417-1425. (in Chinese).
[27]	张寅, 朱桂熠, 施天俊, 等. 基于特征融合与注意力的遥感图像小目标检测[J]. 光学学报,2022,42(24):2415001. ZHANG Y, ZHU G Y, SHI T J, et al. Small object detection in remote sensing images based on feature fusion and attention[J]. Acta Optica Sinica, 2022, 42(24): 2415001. (in Chinese).
[28]	张印辉, 庄宏, 何自芬, 等. 氨气泄漏混洗自注意力轻量化红外检测[J]. 中国光学(中英文),2023,16(3):607-619. ZHENG Y H, ZHUANG H, HE Z F, et al. Lightweight infrared detection of ammonia leakage using shuffle and self-attention[J]. Chinese Optics, 2023, 16(3): 607-619. (in Chinese).
[29]	刘彦磊, 李孟喆, 王宣宣. 轻量型YOLOv5s车载红外图像目标检测[J]. 中国光学(中英文),2023,16(5):1045-1055. LIU Y L, LI M ZH, WANG X X. Lightweight YOLOv5s vehicle infrared image target detection[J]. Chinese Optics, 2023, 16(5): 1045-1055. (in Chinese).
[30]	VILLARRAGA-GÓMEZ H, SMITH S T. Effect of geometric magnification on dimensional measurements with a metrology-grade X-ray computed tomography system[J]. Precision Engineering, 2022, 73: 488-503.