基于一维卷积的生产线冷态重轨表面缺陷快速检测

doi:10.12068/j.issn.1005-3026.2021.02.018

东北大学学报（自然科学版） ›› 2021, Vol. 42 ›› Issue (2): 276-281.DOI: 10.12068/j.issn.1005-3026.2021.02.018

基于一维卷积的生产线冷态重轨表面缺陷快速检测

张德富，宋克臣，牛孟辉，颜云辉

(东北大学机械工程与自动化学院，辽宁沈阳110819)

收稿日期:2020-07-24 修回日期:2020-07-24 接受日期:2020-07-24 发布日期:2021-03-05
通讯作者: 张德富
作者简介:张德富(1993-)，男，河北张家口人，东北大学博士研究生; 颜云辉(1960-)，男，江苏丹阳人，东北大学教授，博士生导师.
基金资助:
国家重点研发计划项目(2017YFB0304200)；国家自然科学基金资助项目(51805078)；中央高校基本科研业务费专项资金资助项目(N2003021).

Rapid Detection of Cold Heavy Rail Surface Defects of Production Line Based on One-Dimensional Convolution Network

ZHANG De-fu， SONG Ke-chen， NIU Meng-hui， YAN Yun-hui

School of Mechanical Engineering & Automation， Northeastern University， Shenyang 110819， China.

Received:2020-07-24 Revised:2020-07-24 Accepted:2020-07-24 Published:2021-03-05
Contact: YAN Yun-hui
About author:-
Supported by:
-

摘要/Abstract

摘要： 采用直观、高效的基于机器视觉的检测方式对生产线冷态重轨表面缺陷进行自动化检测.以彩色双目线阵相机作为采集传感器获取深度信息和RGB信息.深度信息用于缺陷快速检测，RGB信息及深度信息用于缺陷分割.然后，提出一个基于一维卷积网络的深度网络用于缺陷快速检测.该网络主要包括基于一维卷积网络的特征提取器，由全连接层和Dropout层组成的分类器，以及加入尺寸先验的滤波器.为了验证所提出的网络性能，本文搭建了数据采集平台并对重轨样件进行了数据采集.实验结果表明，本文的快速检测网络在采集的数据上缺陷级检测率为100%，误检率为35%，优于对比网络.

关键词: 生产线冷态重轨;表面缺陷;机器视觉;深度信息;一维卷积网络

Abstract: An intuitive and efficient method based on machine vision was applied to the automatic detection of cold heavy rail surface defects of production line. Color binocular linear scan camera gathered the depth information and RGB information. Depth information was employed for the rapid detection of defects， and for defect segmentation combined with RGB information. Then a deep learning network was proposed for the rapid detection of defects. The network mainly includes a feature extractor based on one-dimensional convolution network， a classifier composed of full connection layers and dropout layers， and a filter with size prior. Finally， a data acquisition platform was setup and the data of heavy rail samples were collected for the verification of network performance. The results show that the network proposed performs well. The defect-level detection rate is 100% and the false detection rate is 35% on the collected data， which is better than that of the compared networks.

Key words: cold heavy rail of production line; surface defect; machine vision; depth information; one-dimensional convolution network

中图分类号:

TP391.4

张德富，宋克臣，牛孟辉，颜云辉. 基于一维卷积的生产线冷态重轨表面缺陷快速检测[J]. 东北大学学报（自然科学版）, 2021, 42(2): 276-281.

ZHANG De-fu， SONG Ke-chen， NIU Meng-hui， YAN Yun-hui. Rapid Detection of Cold Heavy Rail Surface Defects of Production Line Based on One-Dimensional Convolution Network[J]. Journal of Northeastern University(Natural Science), 2021, 42(2): 276-281.

参考文献

[1]Yu H，Li Q，Tan Y，et al.A coarse-to-fine model for rail surface defect detection［J］.IEEE Transactions on Instrumentation and Measurement， 2019，68(3):656-666.
[2]张银花，周清跃，陈朝阳，等.中国高速铁路用钢轨的质量现状及分析［J］.钢铁，2011，46(12):1-14.(Zhang Yin-hua，Zhou Qing-yue，Chen Zhao-yang，et al.Quality status and analysis of the rail for high-speed railway in China［J］.Iron and Steel，2011，46(12):1-14.)
[3]Gao B，Bai L，Woo W L，et al.Automatic defect identification of eddy current pulsed thermography using single channel blind source separation［J］.IEEE Transactions on Instrumentation and Measurement，2014，63(4):913-922.
[4]Zhu J，Min Q，Wu J，et al.Probability of detection for eddy current pulsed thermography of angular defect quantification［J］.IEEE Transactions on Industrial Informatics，2018，14(12):5658-5666.
[5]Li Y，Trinh H，Haas N，et al.Rail component detection，optimization，and assessment for automatic rail track inspection［J］.IEEE Transactions on Intelligent Transportation Systems，2014，15(2):760-770.
[6]Cruz F C，Filho E F S，Albuquerque M C，et al.Efficient feature selection for neural network based detection of flaws in steel welded joints using ultrasound testing［J］.Ultrasonics，2017，73:1-8.
[7]Gan J R，Li Q Y，Wang J Z，et al.A hierarchical extractor -based visual rail surface inspection system［J］.IEEE Sensors Journal，2017，17(23):7935-7944.
[8]Li Q，Ren S.A visual detection system for rail surface defects［J］.IEEE Transactions on Systems，Man，and Cybernetics，Part C(Applications and Reviews)，2012，42(6):1531-1542.
[9]Jin X T，Wang Y N，Zhang H，et al.DM-RIS:deep multimodel rail inspection system with improved MRF-GMM and CNN［J］.IEEE Transactions on Instrumentation and Measurement，2020，69(4):1051-1065.
[10]Zhang H，Jin X，Wu Q M J，et al.Automatic visual detection system of railway surface defects with curvature filter and improved Gaussian mixture model［J］.IEEE Transactions on Instrumentation and Measurement，2018，67(7):1593-1608.
[11]Zhou P，Xu K，Wang D D.Rail profile measurement based on line-structured light vision［J］.IEEE Access，2018，6:16423-16431.
[12]徐科，杨朝霖，周鹏，等.基于激光线光源的钢轨表面缺陷三维检测方法［J］.机械工程学报，2010，46(8):1-5.(Xu Ke，Yang Chao-lin，Zhou Peng，et al.3D detection technique of surface defects for steel rails based on linear lasers［J］.Journal of Mechanical Engineering，2010，46(8):1-5.)
[13]张华，赵碧霞，刘桂华，等.双目结构光的钢轨表面缺陷检测系统设计［J］.自动化仪表，2018，39(4):92-95.(Zhang Hua，Zhao Bi-xia，Liu Gui-hua，et al.Design of rail surface defect detection system based on binocular structure light［J］.Process Automation Instrumentation，2018，39(4):92-95.)
[14]金列俊，詹建明，陈俊华，等.基于一维卷积神经网络的钻杆故障诊断［J］.浙江大学学报(工学版)，2020，54(3):467-474.(Jin Lie-jun，Zhan Jian-ming，Chen Jun-hua，et al.Drill pipe fault diagnosis method based on one-dimensional convolutional neural network［J］.Journal of Zhejiang University(Engineering Science)，2020，54(3):467-474.)
[15]谢胜龙，张为民，鲁玉军，等.基于一维卷积神经网络的房颤智能诊断方法研究［J］.计量学报，2020，41(5):620-626.(Xie Sheng-long，Zhang Wei-min，Lu Yu-jun，et al.Research on intelligent diagnosis of atrial fibrillation based on one-dimensional convolution neural network［J］.Acta Metrologica Sinica，2020，41(5):620-626.)

基于一维卷积的生产线冷态重轨表面缺陷快速检测

Rapid Detection of Cold Heavy Rail Surface Defects of Production Line Based on One-Dimensional Convolution Network

RichHTML

PDF (PC)

可视化

摘要/Abstract

引用本文

使用本文

参考文献

相关文章 15

编辑推荐

Metrics

本文评价

[1]	丁其川，王力，刘成. 融合长距离信道注意力与病理特征的肺结节分类[J]. 东北大学学报（自然科学版）, 2023, 44(4): 476-485.
[2]	杨丹，刘国如，任梦成，裴宏杨. 多尺度卷积核U-Net模型的视网膜血管分割方法[J]. 东北大学学报（自然科学版）, 2021, 42(1): 7-14.
[3]	原培新，陈鼎夫. 双能X射线高动态范围安检图像压缩算法[J]. 东北大学学报（自然科学版）, 2021, 42(1): 96-101.
[4]	张田，田勇，王子，王昭东. 基于清晰度评价的自适应阈值图像分割法[J]. 东北大学学报:自然科学版, 2020, 41(9): 1231-1238.
[5]	徐久强，张金鹏，贾玉其，邵建新. 基于集成学习的束支传导阻滞识别方法[J]. 东北大学学报:自然科学版, 2020, 41(3): 321-326.
[6]	徐礼胜，宋代远，刘文彦，王瑜璠. 用不同外周动脉波形重建主动脉波形的对比研究[J]. 东北大学学报:自然科学版, 2020, 41(2): 188-192.
[7]	冯宝，李昌林，李智，刘壮盛. 基于活动轮廓模型和影像组学的乳腺癌LVI状态预测[J]. 东北大学学报:自然科学版, 2020, 41(2): 193-199.
[8]	石春鹤，吴成东. 基于加权稀疏非负矩阵分解的车脸识别算法[J]. 东北大学学报:自然科学版, 2019, 40(10): 1376-1381.
[9]	张云洲，郑瑞，暴吉宁，朱尚栋. 基于多个相关滤波器的行人跟踪尺度算法[J]. 东北大学学报:自然科学版, 2019, 40(9): 1228-1234.
[10]	陈红，于晓升，吴成东，孙鹏. 参数化水平集活动轮廓模型的快速图像分割算法[J]. 东北大学学报:自然科学版, 2019, 40(1): 6-10.
[11]	张良钰，边天元，徐礼胜. 冠脉狭窄对外周与中心动脉间传递函数的影响[J]. 东北大学学报:自然科学版, 2018, 39(12): 1702-1707.
[12]	赵海，李雄峰，朱宏博，王彬. 基于分形维数特征的肺结节形状建模[J]. 东北大学学报:自然科学版, 2018, 39(11): 1545-1551.
[13]	窦圣昶，赵海，朱宏博，王彬. 修正与融合的肺部分割算法[J]. 东北大学学报:自然科学版, 2018, 39(11): 1562-1566.
[14]	徐礼胜，张书琪，牛潇，徐阳. 基于全卷积网络的左心室射血分数自动检测[J]. 东北大学学报:自然科学版, 2018, 39(11): 1572-1576.
[15]	齐林，吕旭阳，杨本强，徐礼胜. 基于全卷积网络迁移学习的左心室内膜分割[J]. 东北大学学报:自然科学版, 2018, 39(11): 1577-1582.