锻压技术

基于图像识别的轧辊表面破损检测系统

英文标题：Roll surface damage detection system based on image recognition

作者：杨晋玲1 段牧忻2

分类号：TP271

出版年,卷(期):页码：2021,46(6):225-230

摘要：

考虑到轧辊是轧机的重要组成部件，轧辊的质量直接影响轧机成形产品的质量，研究了基于图像识别的轧辊表面破损检测系统。利用高清摄像机采集轧辊图像；使用滤波去噪算法去除图像中的干扰因素；使用图像分割技术分割图像，以提高图像的识别效率；使用模板匹配的方法进行破损的识别；使用拉普拉斯方法对图像进行锐化处理，从而使得轧辊破损区域的边缘更加清晰，更容易被显示出来；结合采集的图像特征形成专家诊断库。对于新采集到的待识别图像，通过与专家诊断库的数据进行对比，即可自动判别其为何种破损类型。实验表明，研究的检测系统对于腐蚀、划痕或裂纹、剥落这3大类轧辊的破损情况均有较好的检测效果。

The roll is an important component of rolling mill, and the quality of roll directly affects the quality of the products formed by rolling mill. Therefore, the roll surface damage detection system based on image recognition was studied by using high-definition camera to collect roll image, using filter denoising algorithm to remove the interference factors in the image, using image segmentation technology to segment the image in order to improve the efficiency of image recognition, using template matching method to identify the damage, using Laplace method to sharpen the image so that the edge of the damaged area for roll was clearer and easier to be displayed, and combining with the collected image features to form the expert diagnosis database. Furthermore, for the newly collected image to be identified, the damage type was automatically identified by comparing with the data of expert database. The experimental results show that the detection system has good detection effect for three types of roll damage such as corrosion, scratch or crack and spalling.

基金项目：

国家自然科学基金资助项目（61525108）

作者简介：杨晋玲（1992-），女，硕士研究生 E-mail：yjl9212@126.com 通信作者：段牧忻（1969-），女，本科，高级工程师 E-mail：dwwqqq2021@126.com

参考文献：

[1]李有智, 季业益, 陆宝山,等. 轧辊辊缝差和轧机组装间隙对精轧钢带尾板侧偏的影响[J]. 锻压技术,2021,46(1):116-125.

Li Y Z, Ji Y Y, Lu B S, et al. Influence of roll gap difference and mill assembly gap on side deviation of finished steel strip tail plate [J]. Forging & Stamping Technology,2021,46(1):116-125.

[2]黄鑫, 吴君三, 朱乾皓,等.70Cr3NiMo钢轧辊探伤不合格原因分析及改善措施[J]. 锻压技术,2019,44(10):20-24.

Huang X, Wu J S, Zhu Q H, et al. Analysis on disqualification reason and improvement measures of steel 70Cr3NiMo roller by ultrasonic inspection [J]. Forging & Stamping Technology,2019,44(10):20-24.

[3]曹建国,黄小海,赵秋芳,等.板带轧机通用变凸度板形控制技术[J].中南大学学报：自然科学版,2020,51(10):2772-2781.

Cao J G, Huang X H, Zhao Q F, et al. Universal variable crown technology for strip profile control in wide strip rolling mills [J]. Journal of Central South University：Science and Technology,2020,51(10):2772-2781.

[4]徐科,周鹏,贺笛,等.机器视觉技术及在钢铁生产中的应用[A].第十二届中国钢铁年会论文集——大会特邀报告&分会场特邀报告[C]. 北京:中国金属学会,2019.

Xu K, Zhou P, He D, et al. Machine vision technology and its application in iron and steel production [A]. Proceedings of the 12th China Iron and Steel Annual meeting-invited report of the General Assembly and Invited Report of the Branch Venue[C]. Beijing:China Metal Society, 2019.

[5]梁颖,詹光曹,徐科.基于二值化赋范梯度的中厚板表面缺陷检测[J].表面技术, 2019,48(10):336-341.

Liang Y, Zhan G C, Xu K. Surface defect detection of medium and heavy plates based on binarized normed gradients [J]. Surface Technology,2019,48(10):336-341.

[6]饶静,杨立庆,王超海,等.热连轧精轧机工作辊表面质量问题分析与改进[J].特钢技术,2019,25(4):56-59.

Rao J, Yang L Q, Wang C H, et al. Analysis and improvement of surface quality of finishing mill working roll [J]. Special Steel Technology,2019,25(4):56-59.

[7]于浩,于文禹. 利用无损检测方法探测轧辊缺陷的可靠性[J].无损探伤,2018,42(6):39-41.

Yu H, Yu W Y. Reliability of detecting roll defects by non-destructive testing method [J]. Nondestructive Inspection,2018,42(6):39-41.

[8]吴昆鹏,石杰.基于孪生网络的带钢表面周期性缺陷检测方法[J].冶金自动化,2020,44(6):93-98.

Wu K P, Shi J. Method for periodic defect detection of strip surface based on siamese network [J]. Metallurgical Industry Automation,2020,44(6):93-98.

[9]李永婷,夏琴香,肖刚锋,等. 基于机器视觉的锥形旋压件起皱缺陷在线检测方法[J]. 锻压技术,2019,44(1):134-141.

Li Y T, Xia Q X, Xiao G F, et al. On-line detection method of wrinkling defects of conical spinning parts based on machine vision [J]. Forging & Stamping Technology, 2019,44(1):134-141.

[10]李立新, 黄英钢, 张葵. 轧辊磨损过程中的形貌图像特征及分形[J].钢铁, 2015, 50(4):95-100.

Li L X, Huang Y G, Zhang K. Morphology image features and fractals during roll wear process[J]. Iron & Steel, 2015, 50(4): 95-100.

[11]肖艳军,齐浩,周围,等.锂电池极片轧机轧辊表面缺陷检测与识别[J].电子测量与仪器学报,2019,33(10):148-156.

Xiao Y J, Qi H, Zhou W, et al. Detection and recognition of roll surface defects in lithium battery pole rolling mill [J]. Journal of Electronic Measurement and Instrumentation, 2019,33 (10):148-156.

[12]蒋树强,闵巍庆,王树徽. 面向智能交互的图像识别技术综述与展望[J]. 计算机研究与发展, 2016, 53(1): 113-122.

Jiang S Q, Min W Q, Wang S H. Overview and prospect of image recognition technology for intelligent interaction [J]. Journal of Computer Research and Development, 2016, 53 (1): 113-122.

[13]金洋,王日新,徐敏强. 基于状态记忆的航天器自主故障诊断方法[J]. 系统工程与电子技术, 2015, 37(6): 1452-1458.

Jin Y, Wang R X, Xu M Q. Autonomous fault diagnosis method for spacecraft based on state memory [J]. Systems Engineering and Electronics, 2015, 37 (6): 1452-1458.

服务与反馈：

【文章下载】【加入收藏】