EUVIMEDEUVIMED entdecken
Open Access · CC BY 4.0

YOLO-Weld: A Modified YOLOv5-Based Weld Feature Detection Network for Extreme Weld Noise

Ang Gao iD, Zhuoxuan Fan iD, Anning Li, Qiaoyue Le, Dongting Wu iD, Fuxin Du iD

Sensors · 16.06.2023 · 23 (12) · 5640

DOI: 10.3390/s23125640 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Sensors, online erschienen am 16.06.2023. Verfasst von Ang Gao und 5 weiteren Autor:innen. Beteiligte Einrichtung: School of Mechanical Engineering, Shandong University, Jinan 250061, China. Hauptthemen laut MEDLINE: Welding. Die Arbeit wurde bisher 34-mal zitiert (Crossref) und nennt 45 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0). Gefördert von: Key Research and Development Program of Shandong Province.

AbstractSchlagwörterBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Weld feature point detection is a key technology for welding trajectory planning and tracking. Existing two-stage detection methods and conventional convolutional neural network (CNN)-based approaches encounter performance bottlenecks under extreme welding noise conditions. To better obtain accurate weld feature point locations in high-noise environments, we propose a feature point detection network, YOLO-Weld, based on an improved You Only Look Once version 5 (YOLOv5). By introducing the reparameterized convolutional neural network (RepVGG) module, the network structure is optimized, enhancing detection speed. The utilization of a normalization-based attention module (NAM) in the network enhances the network’s perception of feature points. A lightweight decoupled head, RD-Head, is designed to improve classification and regression accuracy. Furthermore, a welding noise generation method is proposed, increasing the model’s robustness in extreme noise environments. Finally, the model is tested on a custom dataset of five weld types, demonstrating better performance than two-stage detection methods and conventional CNN approaches. The proposed model can accurately detect feature points in high-noise environments while meeting real-time welding requirements. In terms of the model’s performance, the average error of detecting feature points in images is 2.100 pixels, while the average error in the world coordinate system is 0.114 mm, sufficiently meeting the accuracy needs of various practical welding tasks.

Schlagwörter

MeSH-Schlagwörter (MEDLINE)

WeldingCultureExtreme EnvironmentsNeural Networks, ComputerTechnology

Stichwörter der Autor:innen

convolutional neural networkfeature extractionintelligent robotic weldinglaser visual sensorstructured-light vision

Autor:innen und Institutionen

  • School of Mechanical Engineering, Shandong University, Jinan 250061, China.
  • Key Laboratory of High-Efficiency and Clean Mechanical Manufacture, Shandong University, Ministry of Education, Jinan 250061, China.
  • Key Laboratory of Liquid-Solid Structural Evolution and Processing of Materials, Shandong University, Ministry of Education, Jinan 250061, China.
  • Engineering Research Center of Intelligent Unmanned System, Ministry of Education, Jinan 250061, China.

ORCID-Profile: Ang Gao, Zhuoxuan Fan, Dongting Wu, Fuxin Du

Förderung

  • Key Research and Development Program of Shandong Province

Zitieren

APA

Gao, A., Fan, Z., Li, A., Le, Q., Wu, D., & Du, F. (2023). YOLO-Weld: A Modified YOLOv5-Based Weld Feature Detection Network for Extreme Weld Noise. Sensors, 23(12), 5640. https://doi.org/10.3390/s23125640

Vancouver

Gao A, Fan Z, Li A, Le Q, Wu D, Du F. YOLO-Weld: A Modified YOLOv5-Based Weld Feature Detection Network for Extreme Weld Noise. Sensors. 2023;23(12):5640. doi:10.3390/s23125640

BibTeX
@article{Gao2023,
  author = {Gao, Ang and Fan, Zhuoxuan and Li, Anning and Le, Qiaoyue and Wu, Dongting and Du, Fuxin},
  title = {YOLO-Weld: A Modified YOLOv5-Based Weld Feature Detection Network for Extreme Weld Noise},
  journal = {Sensors},
  year = {2023},
  volume = {23},
  number = {12},
  pages = {5640},
  doi = {10.3390/s23125640}
}
Kostenlos starten →