基于CNN的带钢表面缺陷检测
|
摘要
随着工业4.0:智能制造理念的深入,推动工业生产制造进一步由自动化向智能化升级势不可挡。作为制造业的支柱性产业,实现钢铁行业的表面缺陷自主检测具有重要意义。针对现有带钢表面缺陷检测过程中识别率低,无法自主检测等问题,本文提出一种基于CNN的带钢表面缺陷检测算法,引入深度学习知识,通过建立CNN模型,制作数据集,实现了对带钢表面缺陷的自动提取与检测。通过实验验证了该算法的有效性,实验结果表明该算法的准确率在99.99%以上,实验过程中未出现误差,能够满足工业生产方面的要求。
With the deepening of the industry 4.0: intelligent manufacturing concept, promoting industrial production and manufacturing further from automation to intelligent upgrade is unstoppable. As a pillar industry of the manufacturing industry, it is of great significance to realize the self-detection of surface defects in the steel industry. Aiming at the problems of low recognition rate and inability to detect independently in the process of existing strip steel surface defect detection, this paper proposes a CNN-based strip steel surface defect detection algorithm, by introducing deep learning knowledge, establishing CNN model and making data sets, the automatic extraction and detection of strip steel surface defects is realized. The effectiveness of the algorithm is verified by experiments. The experimental result shows that the accuracy of the algorithm is above 99.99%, which is basically error-free and can meet the requirements of industrial production.
With the deepening of the industry 4.0: intelligent manufacturing concept, promoting industrial production and manufacturing further from automation to intelligent upgrade is unstoppable. As a pillar industry of the manufacturing industry, it is of great significance to realize the self-detection of surface defects in the steel industry. Aiming at the problems of low recognition rate and inability to detect independently in the process of existing strip steel surface defect detection, this paper proposes a CNN-based strip steel surface defect detection algorithm, by introducing deep learning knowledge, establishing CNN model and making data sets, the automatic extraction and detection of strip steel surface defects is realized. The effectiveness of the algorithm is verified by experiments. The experimental result shows that the accuracy of the algorithm is above 99.99%, which is basically error-free and can meet the requirements of industrial production.
引文
[1] 熊鹰.带钢表面缺陷检测及识别关键技术研究[D].重庆:重庆大学,2016.
[2] Fang Guo,Jiang Zhao,Ping Jiang. Surface Defects Detection of Steel Plate Based on Visual Attention Mechanism. 2017 Chinese Automation Congress (CAC)[C].2017:3174-3177
[3] 刘坤,陈宁纪,张晓怿,等.基于机器视觉的带钢表面缺陷检测系统设计[J].制造业自动化,2015(06):134-137.
[4] 孙梓钧.基于AdaBoost分类器的带钢表面缺陷识别研究[D].河北:河北科技大学,2016.
[5] 王立中,管声启.基于深度学习算法的带钢表面缺陷识别[J].西安工程大学学报,2017(10):669-674.
[6] 吕铁,韩娜.智能制造:全球趋势与中国战略[J].人民论坛(学术前沿),2015,(11).
[7] 王叶.工业4.0背景下传统制造业转型升级新思维与新模式[J].中国设备工程,2017(10).
[8] 郭朝晖.钢铁行业与工业4.0[J].冶金自动化,2015(07):7-11.
[9] 任浩,屈剑锋,柴毅,等.深度学习在故障诊断领域中的研究现状与挑战[J].控制与决策,2017,(08).
[10] Laila Ma'rifatul Azizah, Sitti Fadillah Umayah, Slamet Riyadi, Cahya Damarjati. Nafi Ananda Utama.Deep Learning Implementation using Convolutional Neural Network in Mangosteen Surface Defect Detection[A]//2017 7th IEEE International Conference on Control System[C]. Computing and Engineering (ICCSCE).2017:242-246.
[11] Lidan Shang, Qiushi Yang, Jianing Wang, Shubin Li, Weimin Lei. Detection of Rail Surface Defects Based on CNN Image Recognition and Classification[C]. 2018 20th International Conference on Advanced Communication Technology (ICACT).2018:45-51.
[12] 林懿伦,戴星原,李力,等.人工智能研究的新前线:生成式对抗网络[J].自动化学报,2018,(05).
[13] Kingma, D. P., & Ba, J. L. Adam: a Method for Stochastic Optimization[C]. International Conference on Learning Representations, 2015:1-13.
[2] Fang Guo,Jiang Zhao,Ping Jiang. Surface Defects Detection of Steel Plate Based on Visual Attention Mechanism. 2017 Chinese Automation Congress (CAC)[C].2017:3174-3177
[3] 刘坤,陈宁纪,张晓怿,等.基于机器视觉的带钢表面缺陷检测系统设计[J].制造业自动化,2015(06):134-137.
[4] 孙梓钧.基于AdaBoost分类器的带钢表面缺陷识别研究[D].河北:河北科技大学,2016.
[5] 王立中,管声启.基于深度学习算法的带钢表面缺陷识别[J].西安工程大学学报,2017(10):669-674.
[6] 吕铁,韩娜.智能制造:全球趋势与中国战略[J].人民论坛(学术前沿),2015,(11).
[7] 王叶.工业4.0背景下传统制造业转型升级新思维与新模式[J].中国设备工程,2017(10).
[8] 郭朝晖.钢铁行业与工业4.0[J].冶金自动化,2015(07):7-11.
[9] 任浩,屈剑锋,柴毅,等.深度学习在故障诊断领域中的研究现状与挑战[J].控制与决策,2017,(08).
[10] Laila Ma'rifatul Azizah, Sitti Fadillah Umayah, Slamet Riyadi, Cahya Damarjati. Nafi Ananda Utama.Deep Learning Implementation using Convolutional Neural Network in Mangosteen Surface Defect Detection[A]//2017 7th IEEE International Conference on Control System[C]. Computing and Engineering (ICCSCE).2017:242-246.
[11] Lidan Shang, Qiushi Yang, Jianing Wang, Shubin Li, Weimin Lei. Detection of Rail Surface Defects Based on CNN Image Recognition and Classification[C]. 2018 20th International Conference on Advanced Communication Technology (ICACT).2018:45-51.
[12] 林懿伦,戴星原,李力,等.人工智能研究的新前线:生成式对抗网络[J].自动化学报,2018,(05).
[13] Kingma, D. P., & Ba, J. L. Adam: a Method for Stochastic Optimization[C]. International Conference on Learning Representations, 2015:1-13.