EUVIMEDEUVIMED entdecken
Open Access · CC BY 4.0

Attention-Enhanced Semantic Segmentation for Substation Inspection Robot Navigation

Changqing Cai iD, Yongkang Yang, Kaiqiao Tian iD, Yuxin Yan, Kazuyuki Kobayashi iD, Ka C Cheok

Sensors · 09.10.2025 · 25 (19) · 6252

DOI: 10.3390/s25196252 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Sensors, online erschienen am 09.10.2025. Verfasst von Changqing Cai und 5 weiteren Autor:innen. Beteiligte Einrichtung: College of Electrical and Information Engineering, Changchun Institute of Technology, 395 Kuan Ping Road, Changchun 130103, China. Die Arbeit wurde bisher 2-mal zitiert (Crossref) und nennt 29 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0). Gefördert von: Key Project Jilin Province Science and Technology Development Plan.

AbstractSchlagwörterBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Outdoor substations present complex conditions such as uneven terrain, strong illumination variations, and frequent occlusions, which pose significant challenges for autonomous robotic inspection. To address these issues, we develop an embedded inspection robot that integrates attention-enhanced semantic segmentation with GPS-assisted navigation for reliable operation. A lightweight DeepLabV3+ model is improved with ECA-SimAM and CBAM attention modules and further extended with a GPS-guided attention component that incorporates coarse location priors to refine feature focus and improve boundary recognition under challenging lighting and occlusion. The segmentation outputs are used to generate real-time road masks and navigation lines via center-of-mass and least-squares fitting, while RTK-GPS provides global positioning and triggers waypoint-based behaviors such as turning and stopping. Experimental results show that the proposed method achieves 85.26% mean IoU and 89.45% mean pixel accuracy, outperforming U-Net, PSPNet, HRNet, and standard DeepLabV3+. Deployed on an embedded platform and validated in real substations, the system demonstrates both robustness and scalability for practical infrastructure inspection tasks.

Schlagwörter

Stichwörter der Autor:innen

DeepLabV3+attention mechanismautonomous inspection robotmultimodal perceptionnavigation line fittingsemantic segmentation

Autor:innen und Institutionen

  • College of Electrical and Information Engineering, Changchun Institute of Technology, 395 Kuan Ping Road, Changchun 130103, China.
  • Electrical and Computer Engineering, Oakland University, Rochester, MI 48309, USA.
  • Department of Advanced Sciences, Hosei University, Tokyo 184-8584, Japan.

ORCID-Profile: Changqing Cai, Kaiqiao Tian, Kazuyuki Kobayashi

Förderung

  • Key Project Jilin Province Science and Technology Development Plan

Zitieren

APA

Cai, C., Yang, Y., Tian, K., Yan, Y., Kobayashi, K., & Cheok, K. C. (2025). Attention-Enhanced Semantic Segmentation for Substation Inspection Robot Navigation. Sensors, 25(19), 6252. https://doi.org/10.3390/s25196252

Vancouver

Cai C, Yang Y, Tian K, Yan Y, Kobayashi K, Cheok KC. Attention-Enhanced Semantic Segmentation for Substation Inspection Robot Navigation. Sensors. 2025;25(19):6252. doi:10.3390/s25196252

BibTeX
@article{Cai2025,
  author = {Cai, Changqing and Yang, Yongkang and Tian, Kaiqiao and Yan, Yuxin and Kobayashi, Kazuyuki and Cheok, Ka C},
  title = {Attention-Enhanced Semantic Segmentation for Substation Inspection Robot Navigation},
  journal = {Sensors},
  year = {2025},
  volume = {25},
  number = {19},
  pages = {6252},
  doi = {10.3390/s25196252}
}
Kostenlos starten →