EUVIMEDEUVIMED entdecken
Crossref

PhysFusion: A Transformer-Based Dual-Stream Radar and Vision Fusion Framework for Open Water-Surface Object Detection

Yuting Wan, Liguo Sun, Jiuwu Hao, Zao Zhang, Pin Lv

IEEE Sensors Journal · 15.07.2026 · 26 (14) · 21369-21380

DOI: 10.1109/jsen.2026.3697429 ↗

Kurz gefasst

Fachartikel in der Zeitschrift IEEE Sensors Journal, erschienen am 15.07.2026. Verfasst von Yuting Wan und 4 weiteren Autor:innen. Die Arbeit wurde bisher 2-mal zitiert (Crossref) und nennt 46 Literaturangaben.

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Detecting water-surface targets for autonomous surface vehicle (ASVs) is challenging due to nonstationary wave clutter, specular reflections, and frequent long-range observations, where small or low-profile objects exhibit weak appearance cues. Although 4-D millimeter-wave radar can complement cameras under degraded illumination, maritime radar point clouds are typically sparse and intermittent, and reflectivity-related attributes often present heavy-tailed variations under scattering and multipath, making it difficult for conventional fusion designs to exploit radar cues effectively and maintain stable predictions across frames. We propose PhysFusion, a physics-informed radar–image detection framework tailored for water-surface perception. Instead of relying on direct fusion of sparse radar returns and dense visual features, PhysFusion improves the multimodal detection from three complementary aspects: radar representation learning under cluttered and heavy-tailed radar observations, query-level radar-guided cross-modal interaction for more effective fusion with multiscale visual features, and lightweight temporal aggregation for stable frame-to-frame prediction. This design enables more reliable radar cue utilization and more consistent multimodal detection in challenging water-surface environments. Experiments on WaterScenes and FLOW demonstrate that PhysFusion achieves 61.5% mAP50:95 and 90.3% mAP50 on WaterScenes (with T = 5 radar history) using 5.6M parameters and 12.5G FLOPs and reaches 94.8% mAP50 and 46.2% mAP50:95 on FLOW under the radar–camera setting. Ablation studies further verify the effectiveness of the proposed design under a consistent evaluation protocol.

Zitieren

APA

Wan, Y., Sun, L., Hao, J., Zhang, Z., & Lv, P. (2026). PhysFusion: A Transformer-Based Dual-Stream Radar and Vision Fusion Framework for Open Water-Surface Object Detection. IEEE Sensors Journal, 26(14), 21369-21380. https://doi.org/10.1109/jsen.2026.3697429

Vancouver

Wan Y, Sun L, Hao J, Zhang Z, Lv P. PhysFusion: A Transformer-Based Dual-Stream Radar and Vision Fusion Framework for Open Water-Surface Object Detection. IEEE Sensors Journal. 2026;26(14):21369-21380. doi:10.1109/jsen.2026.3697429

BibTeX
@article{Wan2026,
  author = {Wan, Yuting and Sun, Liguo and Hao, Jiuwu and Zhang, Zao and Lv, Pin},
  title = {PhysFusion: A Transformer-Based Dual-Stream Radar and Vision Fusion Framework for Open Water-Surface Object Detection},
  journal = {IEEE Sensors Journal},
  year = {2026},
  volume = {26},
  number = {14},
  pages = {21369--21380},
  doi = {10.1109/jsen.2026.3697429}
}
Kostenlos starten →