EUVIMEDEUVIMED entdecken
Open Access · CC BY 4.0

HSF-DETR: Hyper Scale Fusion Detection Transformer for Multi-Perspective UAV Object Detection

Yi Mao, Haowei Zhang, Rui Li, Feng Zhu, Rui Sun, Pingping Ji

Remote Sensing · 09.06.2025 · 17 (12) · 1997

DOI: 10.3390/rs17121997 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Remote Sensing, online erschienen am 09.06.2025. Verfasst von Yi Mao und 5 weiteren Autor:innen. Die Arbeit wurde bisher 17-mal zitiert (Crossref) und nennt 60 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0). Gefördert von: National Natural Science Foundation of China, Jiangsu Funding Program for Excellent Postdoctoral Talent, Fundamental Research Foundation of National Key Laboratory of Automatic Target Recognition.

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Unmanned aerial vehicle (UAV) imagery detection faces challenges in preserving small object features during multi-level downsampling, handling angle and altitude-dependent variations in aerial scenes, achieving accurate localization in dense environments, and performing real-time detection. To address these limitations, we propose HSF-DETR, a lightweight transformer-based detector specifically designed for UAV imagery. First, we design a hybrid progressive fusion network (HPFNet) as the backbone, which adaptively modulates receptive fields to capture multi-scale information while preserving fine-grained details critical for small object detection. Second, building upon features extracted by HPFNet, we develop MultiScaleNet, which enhances feature representation through dual-layer optimization and cross-domain feature learning, significantly improving the model’s capability to handle complex aerial scenarios with diverse object orientations. Finally, to address spatial–semantic alignment challenges, we devise a position-aware align context and spatial tuning (PACST) module that ensures effective feature calibration through precise alignment and adaptive fusion across scales. This hierarchical architecture is complemented by our novel AdaptDist-IoU loss with dynamic weight allocation, which enhances localization accuracy, particularly in dense environments. Extensive experiments using standard detection metrics (mAP50 and mAP50:95) on the VisDrone2019 test dataset demonstrate that HSF-DETR achieves superior performance with 0.428 mAP50 (+5.4%) and 0.253 mAP50:95 (+4%) when compared with RT-DETR, while maintaining real-time inference (69.3 FPS) on an NVIDIA RTX 4090D GPU with only 15.24M parameters and 63.6 GFLOPs. Further validation across multiple public remote sensing datasets confirms the robust generalization capability of HSF-DETR in diverse aerial scenarios, offering a practical solution for resource-constrained UAV applications where both detection quality and processing speed are crucial.

Förderung

  • National Natural Science Foundation of China
  • Jiangsu Funding Program for Excellent Postdoctoral Talent
  • Fundamental Research Foundation of National Key Laboratory of Automatic Target Recognition
  • Key Laboratory Program

Zitieren

APA

Mao, Y., Zhang, H., Li, R., Zhu, F., Sun, R., & Ji, P. (2025). HSF-DETR: Hyper Scale Fusion Detection Transformer for Multi-Perspective UAV Object Detection. Remote Sensing, 17(12), 1997. https://doi.org/10.3390/rs17121997

Vancouver

Mao Y, Zhang H, Li R, Zhu F, Sun R, Ji P. HSF-DETR: Hyper Scale Fusion Detection Transformer for Multi-Perspective UAV Object Detection. Remote Sensing. 2025;17(12):1997. doi:10.3390/rs17121997

BibTeX
@article{Mao2025,
  author = {Mao, Yi and Zhang, Haowei and Li, Rui and Zhu, Feng and Sun, Rui and Ji, Pingping},
  title = {HSF-DETR: Hyper Scale Fusion Detection Transformer for Multi-Perspective UAV Object Detection},
  journal = {Remote Sensing},
  year = {2025},
  volume = {17},
  number = {12},
  pages = {1997},
  doi = {10.3390/rs17121997}
}
Kostenlos starten →