Deep Learning-Based Weld Contour and Defect Detection from Micrographs of Laser Beam Welded Semi-Finished Products

Research output: Contribution to journal › Article › Research › peer review

Authors

  • Christian Nowroth
  • Tiansheng Gu
  • Jan Grajczak
  • Sarah Nothdurft
  • Jens Twiefel
  • Jörg Hermsdorf
  • Stefan Kaierle
  • Jörg Wallaschek

Research Organisations

External Research Organisations

  • Laser Zentrum Hannover e.V. (LZH)
View graph of relations

Details

Original languageEnglish
Article number4645
JournalApplied Sciences (Switzerland)
Volume12
Issue number9
Publication statusPublished - 5 May 2022

Abstract

Laser beam welding is used in many areas of industry and research. There are many strategies and approaches to further improve the weld seam properties in laser beam welding. Metallography is often needed to evaluate welded seams. Typically, the images are examined and evaluated by experts. The evaluation process qualitatively provides the properties of the welds. Particularly in times when artificial intelligence is being used more and more in processes, the quantization of properties that could previously only be determined qualitatively is gaining importance. In this contribution, we propose to use deep learning to perform semantic segmentation of micrographs of complex weld areas to achieve the automatic detection and quantization of weld seam properties. A semantic segmentation dataset is created containing 282 labeled images. The training process is performed with DeepLabv3+. The trained model achieves a value of around 95% for weld contour detection and 76.88% of mean intersection over union (mIoU).

Keywords

    automatic detection, dataset creation, deep learning, quantization, semantic segmentation, weld defects, weld seam

ASJC Scopus subject areas

Cite this

Deep Learning-Based Weld Contour and Defect Detection from Micrographs of Laser Beam Welded Semi-Finished Products. / Nowroth, Christian; Gu, Tiansheng; Grajczak, Jan et al.
In: Applied Sciences (Switzerland), Vol. 12, No. 9, 4645, 05.05.2022.

Research output: Contribution to journal › Article › Research › peer review

Nowroth, C, Gu, T, Grajczak, J, Nothdurft, S, Twiefel, J, Hermsdorf, J, Kaierle, S & Wallaschek, J 2022, 'Deep Learning-Based Weld Contour and Defect Detection from Micrographs of Laser Beam Welded Semi-Finished Products', Applied Sciences (Switzerland), vol. 12, no. 9, 4645. https://doi.org/10.3390/app12094645
Nowroth, C., Gu, T., Grajczak, J., Nothdurft, S., Twiefel, J., Hermsdorf, J., Kaierle, S., & Wallaschek, J. (2022). Deep Learning-Based Weld Contour and Defect Detection from Micrographs of Laser Beam Welded Semi-Finished Products. Applied Sciences (Switzerland), 12(9), Article 4645. https://doi.org/10.3390/app12094645
Nowroth C, Gu T, Grajczak J, Nothdurft S, Twiefel J, Hermsdorf J et al. Deep Learning-Based Weld Contour and Defect Detection from Micrographs of Laser Beam Welded Semi-Finished Products. Applied Sciences (Switzerland). 2022 May 5;12(9):4645. doi: 10.3390/app12094645
Download
@article{31d6671bb4724fa088d1f23ea877d87a,
title = "Deep Learning-Based Weld Contour and Defect Detection from Micrographs of Laser Beam Welded Semi-Finished Products",
abstract = "Laser beam welding is used in many areas of industry and research. There are many strategies and approaches to further improve the weld seam properties in laser beam welding. Metallography is often needed to evaluate welded seams. Typically, the images are examined and evaluated by experts. The evaluation process qualitatively provides the properties of the welds. Particularly in times when artificial intelligence is being used more and more in processes, the quantization of properties that could previously only be determined qualitatively is gaining importance. In this contribution, we propose to use deep learning to perform semantic segmentation of micrographs of complex weld areas to achieve the automatic detection and quantization of weld seam properties. A semantic segmentation dataset is created containing 282 labeled images. The training process is performed with DeepLabv3+. The trained model achieves a value of around 95% for weld contour detection and 76.88% of mean intersection over union (mIoU).",
keywords = "automatic detection, dataset creation, deep learning, quantization, semantic segmentation, weld defects, weld seam",
author = "Christian Nowroth and Tiansheng Gu and Jan Grajczak and Sarah Nothdurft and Jens Twiefel and J{\"o}rg Hermsdorf and Stefan Kaierle and J{\"o}rg Wallaschek",
note = "Funding Information: Funding: Funded by the Deutsche Forschungsgemeinschaft (DFG, German Research Foundation)— CRC 1153, subproject A3—252662854. The authors would like to thank them for the support.",
year = "2022",
month = may,
day = "5",
doi = "10.3390/app12094645",
language = "English",
volume = "12",
journal = "Applied Sciences (Switzerland)",
issn = "2076-3417",
publisher = "Multidisciplinary Digital Publishing Institute",
number = "9",

}

Download

TY - JOUR

T1 - Deep Learning-Based Weld Contour and Defect Detection from Micrographs of Laser Beam Welded Semi-Finished Products

AU - Nowroth, Christian

AU - Gu, Tiansheng

AU - Grajczak, Jan

AU - Nothdurft, Sarah

AU - Twiefel, Jens

AU - Hermsdorf, Jörg

AU - Kaierle, Stefan

AU - Wallaschek, Jörg

N1 - Funding Information: Funding: Funded by the Deutsche Forschungsgemeinschaft (DFG, German Research Foundation)— CRC 1153, subproject A3—252662854. The authors would like to thank them for the support.

PY - 2022/5/5

Y1 - 2022/5/5

N2 - Laser beam welding is used in many areas of industry and research. There are many strategies and approaches to further improve the weld seam properties in laser beam welding. Metallography is often needed to evaluate welded seams. Typically, the images are examined and evaluated by experts. The evaluation process qualitatively provides the properties of the welds. Particularly in times when artificial intelligence is being used more and more in processes, the quantization of properties that could previously only be determined qualitatively is gaining importance. In this contribution, we propose to use deep learning to perform semantic segmentation of micrographs of complex weld areas to achieve the automatic detection and quantization of weld seam properties. A semantic segmentation dataset is created containing 282 labeled images. The training process is performed with DeepLabv3+. The trained model achieves a value of around 95% for weld contour detection and 76.88% of mean intersection over union (mIoU).

AB - Laser beam welding is used in many areas of industry and research. There are many strategies and approaches to further improve the weld seam properties in laser beam welding. Metallography is often needed to evaluate welded seams. Typically, the images are examined and evaluated by experts. The evaluation process qualitatively provides the properties of the welds. Particularly in times when artificial intelligence is being used more and more in processes, the quantization of properties that could previously only be determined qualitatively is gaining importance. In this contribution, we propose to use deep learning to perform semantic segmentation of micrographs of complex weld areas to achieve the automatic detection and quantization of weld seam properties. A semantic segmentation dataset is created containing 282 labeled images. The training process is performed with DeepLabv3+. The trained model achieves a value of around 95% for weld contour detection and 76.88% of mean intersection over union (mIoU).

KW - automatic detection

KW - dataset creation

KW - deep learning

KW - quantization

KW - semantic segmentation

KW - weld defects

KW - weld seam

UR - http://www.scopus.com/inward/record.url?scp=85131324159&partnerID=8YFLogxK

U2 - 10.3390/app12094645

DO - 10.3390/app12094645

M3 - Article

AN - SCOPUS:85131324159

VL - 12

JO - Applied Sciences (Switzerland)

JF - Applied Sciences (Switzerland)

SN - 2076-3417

IS - 9

M1 - 4645

ER -