EUVIMEDEUVIMED entdecken
Open Access · CC BY 4.0

Bearing-DETR: A Lightweight Deep Learning Model for Bearing Defect Detection Based on RT-DETR

Minggao Liu, Haifeng Wang iD, Luyao Du, Fangsong Ji, Ming Zhang

Sensors · 30.06.2024 · 24 (13) · 4262

DOI: 10.3390/s24134262 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Sensors, online erschienen am 30.06.2024. Verfasst von Minggao Liu und 4 weiteren Autor:innen. Beteiligte Einrichtung: School of Energy and Mining Engineering, Shandong University of Science and Technology, Qingdao 266590, China. Die Arbeit wurde bisher 41-mal zitiert (Crossref) und nennt 26 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0). Gefördert von: Shandong Provincial Natural Science Foundation, China.

AbstractSchlagwörterBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Detecting bearing defects accurately and efficiently is critical for industrial safety and efficiency. This paper introduces Bearing-DETR, a deep learning model optimised using the Real-Time Detection Transformer (RT-DETR) architecture. Enhanced with Dysample Dynamic Upsampling, Efficient Model Optimization (EMO) with Meta-Mobile Blocks (MMB), and Deformable Large Kernel Attention (D-LKA), Bearing-DETR offers significant improvements in defect detection while maintaining a lightweight framework suitable for low-resource devices. Validated on a dataset from a chemical plant, Bearing-DETR outperformed the standard RT-DETR, achieving a mean average precision (mAP) of 94.3% at IoU = 0.5 and 57.5% at IoU = 0.5–0.95. It also reduced floating-point operations (FLOPs) to 8.2 G and parameters to 3.2 M, underscoring its enhanced efficiency and reduced computational demands. These results demonstrate the potential of Bearing-DETR to transform maintenance strategies and quality control across manufacturing environments, emphasising adaptability and impact on sustainability and operational costs.

Schlagwörter

Stichwörter der Autor:innen

bearing defect detectionindustrial efficiencylightweight modelreal-time detection transformer (RT-DETR)

Autor:innen und Institutionen

  • School of Energy and Mining Engineering, Shandong University of Science and Technology, Qingdao 266590, China.
  • School of Information Science and Engineering, Linyi University, Linyi 276002, China.
  • School of Automation, Wuhan University of Technology, Wuhan 430070, China.

ORCID-Profile: Haifeng Wang

Förderung

  • Shandong Provincial Natural Science Foundation, China

Zitieren

APA

Liu, M., Wang, H., Du, L., Ji, F., & Zhang, M. (2024). Bearing-DETR: A Lightweight Deep Learning Model for Bearing Defect Detection Based on RT-DETR. Sensors, 24(13), 4262. https://doi.org/10.3390/s24134262

Vancouver

Liu M, Wang H, Du L, Ji F, Zhang M. Bearing-DETR: A Lightweight Deep Learning Model for Bearing Defect Detection Based on RT-DETR. Sensors. 2024;24(13):4262. doi:10.3390/s24134262

BibTeX
@article{Liu2024,
  author = {Liu, Minggao and Wang, Haifeng and Du, Luyao and Ji, Fangsong and Zhang, Ming},
  title = {Bearing-DETR: A Lightweight Deep Learning Model for Bearing Defect Detection Based on RT-DETR},
  journal = {Sensors},
  year = {2024},
  volume = {24},
  number = {13},
  pages = {4262},
  doi = {10.3390/s24134262}
}
Kostenlos starten →