EUVIMEDEUVIMED entdecken
Open Access · CC BY 4.0

An Enhanced YOLO Framework for Accurate Small-Target Cable Defect Detection

Guangyi Yang, Fan Zhou, Biaofeng Di, Xiaomin Wu, Kai Hu

Electronics · 07.05.2026 · 15 (10) · 1979

DOI: 10.3390/electronics15101979 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Electronics, online erschienen am 07.05.2026. Verfasst von Guangyi Yang und 4 weiteren Autor:innen. Die Arbeit wurde bisher 2-mal zitiert (Crossref) und nennt 24 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0). Gefördert von: Major Scientific and Technological Achievements Industrialization Project of Heilongjiang Province, Natural Science Foundation of Heilongjiang Province.

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Reliable detection of cable surface defects is important for the safety of power transmission systems, but existing methods still face difficulties in small-target representation, multi-scale variation, and complex industrial backgrounds. To address these issues, this paper proposes YOLO-ESBD, an enhanced detection framework based on YOLO11, which focuses on the coordinated integration and optimization of multiple effective modules. Unlike simple module stacking, the method is designed as a coordinated optimization across feature extraction, feature preservation, and prediction stages. EMA is introduced to enhance feature discrimination and suppress background noise. SPD-Conv is used to preserve fine-grained spatial information for small-defect detection. BiFPN improves multi-scale feature fusion, and DyHead enables adaptive detection under varying defect scales and distributions. Experiments on a real industrial dataset show that YOLO-ESBD achieves an mAP@0.5 of 94.2%, outperforming representative baseline methods. Deployment results on edge devices demonstrate real-time performance, reaching up to 86.2 FPS on RK3588 under INT8 quantization, with stable performance across different platforms. Overall, the proposed method achieves a balanced trade-off between accuracy and efficiency, and its improvements come from a coordinated design rather than independent module stacking, making it suitable for industrial edge deployment scenarios.

Förderung

  • Major Scientific and Technological Achievements Industrialization Project of Heilongjiang Province
  • Natural Science Foundation of Heilongjiang Province

Zitieren

APA

Yang, G., Zhou, F., Di, B., Wu, X., & Hu, K. (2026). An Enhanced YOLO Framework for Accurate Small-Target Cable Defect Detection. Electronics, 15(10), 1979. https://doi.org/10.3390/electronics15101979

Vancouver

Yang G, Zhou F, Di B, Wu X, Hu K. An Enhanced YOLO Framework for Accurate Small-Target Cable Defect Detection. Electronics. 2026;15(10):1979. doi:10.3390/electronics15101979

BibTeX
@article{Yang2026,
  author = {Yang, Guangyi and Zhou, Fan and Di, Biaofeng and Wu, Xiaomin and Hu, Kai},
  title = {An Enhanced YOLO Framework for Accurate Small-Target Cable Defect Detection},
  journal = {Electronics},
  year = {2026},
  volume = {15},
  number = {10},
  pages = {1979},
  doi = {10.3390/electronics15101979}
}
Kostenlos starten →