正在加载图片...
工程科学学报,第37卷,增刊1:18-23,2015年5月 Chinese Journal of Engineering,Vol.37,Suppl.1:18-23,May 2015 DOI:10.13374/j.issn2095-9389.2015.s1.004:http://journals.ustb.edu.cn 基于多激光线的钢轨表面缺陷在线检测方法 周鹏四,徐科,张春阳,杨朝霖 北京科技大学高效轧制国家工程研究中心,北京100083 ☒通信作者,E-mail:zhoupeng(@nercar..ustb.edu.cn 摘要用多激光线代替单激光线应用于钢轨表面缺陷在线检测,解决了单激光线模式下由于钢轨跳动造成“伪缺陷”的问 题.对相机采集到的多激光线单幅图像进行了中心线提取、差值算法检测、缺陷判别及缺陷图像拼接等处理,实现了缺陷的 在线检测,并获得缺陷的完整区域图像.该方法利用多激光线单幅图像光带间的互相关和自相关深度信息提取图像的特征 区域进行检测,避免了单激光线模式的深度拼接步骤,从根本上解决了由于钢轨跳动造成的“误检”问题.该方法己经在线应 用于钢轨轨底的表面缺陷检测,相对于单激光线检测具有更高的准确率. 关键词钢轨:表面缺陷:三维检测:结构光 分类号TG156 Oninesurface defect detection for steel rails based on multiine lasers ZHOU Peng,XU Ke,ZHANG Chun-yang,YANG Chao-in National Engineering Research Center for Advanced Rolling Technology,University of Science and Technology Beijing,Beijing 100083,China Corresponding author,E-mail:zhoupeng@nercar.ustb.edu.cn ABSTRACT Multi-line lasers were applied to the on-ine defect detection for steel rails instead of single-ine laser,which solved the problem of "pseudo defects"due to the vibration of rails under the mode of single-ine laser.Images of multi-ine lasers captured by camera were processed to realize the online detection and the accurate defect feature area was acquired with steps of the extraction of strip centers,detection of difference algorithm,discrimination between real and pseudo defects,image mosaic,etc.Images of the defect area were extracted by the cross-correlation and self-correlation information of laser curves,and there was no need for the step of depth mapping under the single-line mode,so the problem of "fault detection"caused by the vibration of rails was solved fundamental- ly.The technique is applied to on-ine detection of defects for the bottom side of steel rails,and the detection rate is much higher than manual detection. KEY WORDS steel rail:surface defect:3-detection:structured lighting 钢轨的表面缺陷是影响钢轨质量的重要因素,因一定的深度,因此可采用基于结构光的二维投影测 此在钢轨出厂之前,需对钢轨表面进行严格检查.目 量方法实现钢轨表面缺陷在线检测.文献2]应用 前,在生产过程中对钢轨进行表面检查的主要方式是 单激光线检测方法和深度拼接算法实现钢轨表面深 人工目测,存在着效率低、劳动强度大、漏检严重等问 度缺陷的在线实时检测,基本上满足了在线检测的 题,已成为钢轨生产的“瓶颈” 要求 表面缺陷在线检测技术可解决人工目测法存在 然而,单激光线的应用,受生产线上钢轨跳动影 的问题,目前应用于钢轨的表面缺陷在线检测主要 响,不同图像间的光带位置会发生偏移、扭转,致使深 采用机器视觉技术▣.由于钢轨表面缺陷往往具有 度拼接过程产生许多“伪缺陷”,影响了检测准确率 收稿日期:201501-20 基金项目:国家科技支撑计划课题(2012BAB19B06):高等学校博士点专项基金资助课题(20120006110033)工程科学学报,第 37 卷,增刊 1: 18--23,2015 年 5 月 Chinese Journal of Engineering,Vol. 37,Suppl. 1: 18--23,May 2015 DOI: 10. 13374 /j. issn2095--9389. 2015. s1. 004; http: / /journals. ustb. edu. cn 基于多激光线的钢轨表面缺陷在线检测方法 周 鹏,徐 科,张春阳,杨朝霖 北京科技大学高效轧制国家工程研究中心,北京 100083  通信作者,E-mail: zhoupeng@ nercar. ustb. edu. cn 摘 要 用多激光线代替单激光线应用于钢轨表面缺陷在线检测,解决了单激光线模式下由于钢轨跳动造成“伪缺陷”的问 题. 对相机采集到的多激光线单幅图像进行了中心线提取、差值算法检测、缺陷判别及缺陷图像拼接等处理,实现了缺陷的 在线检测,并获得缺陷的完整区域图像. 该方法利用多激光线单幅图像光带间的互相关和自相关深度信息提取图像的特征 区域进行检测,避免了单激光线模式的深度拼接步骤,从根本上解决了由于钢轨跳动造成的“误检”问题. 该方法已经在线应 用于钢轨轨底的表面缺陷检测,相对于单激光线检测具有更高的准确率. 关键词 钢轨; 表面缺陷; 三维检测; 结构光 分类号 TG156 On-linesurface defect detection for steel rails based on multi-line lasers ZHOU Peng ,XU Ke,ZHANG Chun-yang,YANG Chao-lin National Engineering Research Center for Advanced Rolling Technology,University of Science and Technology Beijing,Beijing 100083,China  Corresponding author,E-mail: zhoupeng@ nercar. ustb. edu. cn ABSTRACT Multi-line lasers were applied to the on-line defect detection for steel rails instead of single-line laser,which solved the problem of“pseudo defects”due to the vibration of rails under the mode of single-line laser. Images of multi-line lasers captured by camera were processed to realize the online detection and the accurate defect feature area was acquired with steps of the extraction of strip centers,detection of difference algorithm,discrimination between real and pseudo defects,image mosaic,etc. Images of the defect area were extracted by the cross-correlation and self-correlation information of laser curves,and there was no need for the step of depth mapping under the single-line mode,so the problem of“fault detection”caused by the vibration of rails was solved fundamental￾ly. The technique is applied to on-line detection of defects for the bottom side of steel rails,and the detection rate is much higher than manual detection. KEY WORDS steel rail; surface defect; 3-D detection; structured lighting 收稿日期: 2015--01--20 基金项目: 国家科技支撑计划课题 ( 2012BAB19B06) ; 高等学校博士点专项基金资助课题( 20120006110033) 钢轨的表面缺陷是影响钢轨质量的重要因素,因 此在钢轨出厂之前,需对钢轨表面进行严格检查. 目 前,在生产过程中对钢轨进行表面检查的主要方式是 人工目测,存在着效率低、劳动强度大、漏检严重等问 题,已成为钢轨生产的“瓶颈”. 表面缺陷在线检测技术可解决人工目测法存在 的问题,目前应用于钢轨的表面缺陷在线检测主要 采用机器视觉技术[1]. 由于钢轨表面缺陷往往具有 一定的深度,因此可采用基于结构光的二维投影测 量方法实现钢轨表面缺陷在线检测. 文献[2]应用 单激光线检测方法和深度拼接算法实现钢轨表面深 度缺陷的在线实时检测,基本上满足了在线检测的 要求. 然而,单激光线的应用,受生产线上钢轨跳动影 响,不同图像间的光带位置会发生偏移、扭转,致使深 度拼接过程产生许多“伪缺陷”,影响了检测准确率.
向下翻页>>
©2008-现在 cucdc.com 高等教育资讯网 版权所有