EUVIMEDEUVIMED entdecken
Crossref

DCT-ART: A Spatial–Frequency Fusion Network for Robust Ship Wake Detection

Yaxiong Chen, Yang Ding, Yankai Liu, Shili Xiong

IEEE Geoscience and Remote Sensing Letters · 2026 · 23 · 6016805-6016805

DOI: 10.1109/lgrs.2026.3713616 ↗

Kurz gefasst

Fachartikel in der Zeitschrift IEEE Geoscience and Remote Sensing Letters, 2026. Verfasst von Yaxiong Chen und 3 weiteren Autor:innen. Die Arbeit nennt 23 Literaturangaben. Gefördert von: Sanya Science and Technology Special Fund, Project of Sanya Yazhou Bay Science and Technology City, Hainan Province “Nanhai New Star” Technology Innovation Talent Platform Project.

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

The ship wake detection is critical in remote sensing tasks, such as maritime target recognition and behavior analysis. However, wakes are often fragmented and distorted under complex sea conditions. In addition, strong sea clutter overlaps with wake signals, posing severe challenges for conventional feature-driven detection methods. To tackle this, we propose discrete cosine transform (DCT)-aggregated residual transformation (ART), a novel detection network that jointly exploits spatial- and frequency-domain representations. It introduces two key modules: 1) a multibranch ART module with grouped convolutions to extract multiscale spatial features of wakes and 2) a DCT-based channel attention (DCT-CA) module in the frequency domain to highlight discriminative spectral features and suppress background clutter. Experimental results on the SWIM and CGWX datasets demonstrate that DCT-ART achieves superior mean average precision (mAP) and Recall compared to the state-of-the-art methods, while maintaining robustness under low-visibility and high-clutter conditions, thereby improving the accuracy and reliability of wake detection in maritime remote sensing imagery.

Förderung

  • Sanya Science and Technology Special Fund
  • Project of Sanya Yazhou Bay Science and Technology City
  • Hainan Province “Nanhai New Star” Technology Innovation Talent Platform Project

Zitieren

APA

Chen, Y., Ding, Y., Liu, Y., & Xiong, S. (2026). DCT-ART: A Spatial–Frequency Fusion Network for Robust Ship Wake Detection. IEEE Geoscience and Remote Sensing Letters, 23, 6016805-6016805. https://doi.org/10.1109/lgrs.2026.3713616

Vancouver

Chen Y, Ding Y, Liu Y, Xiong S. DCT-ART: A Spatial–Frequency Fusion Network for Robust Ship Wake Detection. IEEE Geoscience and Remote Sensing Letters. 2026;23:6016805-6016805. doi:10.1109/lgrs.2026.3713616

BibTeX
@article{Chen2026,
  author = {Chen, Yaxiong and Ding, Yang and Liu, Yankai and Xiong, Shili},
  title = {DCT-ART: A Spatial–Frequency Fusion Network for Robust Ship Wake Detection},
  journal = {IEEE Geoscience and Remote Sensing Letters},
  year = {2026},
  volume = {23},
  pages = {6016805--6016805},
  doi = {10.1109/lgrs.2026.3713616}
}
Kostenlos starten →