EUVIMEDEUVIMED entdecken
Open Access · CC BY 4.0

DPCNet: A dual path cross perception network for small object detection in UAV imagery

Linfeng Jia iD, Yafeng Zhu iD, Bin Li

PLOS One · 13.03.2026 · 21 (3) · e0344091

DOI: 10.1371/journal.pone.0344091 ↗

Kurz gefasst

Fachartikel in der Zeitschrift PLOS One, online erschienen am 13.03.2026. Verfasst von Linfeng Jia und 2 weiteren Autor:innen. Beteiligte Einrichtung: School of Intelligent Manufacturing and Electrical Engineering, Guangzhou Institute of Science and Technology, Guangzhou, Guangdong Province, China. Hauptthemen laut MEDLINE: Unmanned Aerial Devices, Image Processing, Computer-Assisted. Die Arbeit wurde bisher 2-mal zitiert (Crossref) und nennt 55 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0). Gefördert von: Guangdong Provincial Undergraduate Universities Teaching Quality and Teaching Reform Program—Excellence Program for Talent Training in Intelligent Manufacturing, Guangdong Provincial Key Discipline Construction Project for Scientific Research Capacity Enhancement, University Level First Class Undergraduate Major Program of Guangzhou Institute of Science and Technology—Robotics Engineering.

AbstractSchlagwörterBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Small object detection in unmanned aerial vehicle imagery is challenged by tiny target scales, dense layouts, and cluttered backgrounds that blur fine details and destabilize multiscale representations. We present DPCNet, a single-stage detector that combines dual-path cross perception with deep and shallow feature interaction and a decoupled detection head. The Dual-Path Cross Perception block separates a detail stream and a semantic stream and performs gated bidirectional fusion, preserving edges while enriching context. The Deep and Shallow Feature Interaction block aligns features across levels through dynamic up-sampling and down-sampling and similarity-guided masking, which strengthens cross-scale consistency. The Dual-Path Decoupled Detection Head keeps classification and regression separate yet enables lightweight cross-branch channel and spatial guidance, and bounding-box regression adopts a geometry-sensitive Shape-IoU loss. Experiments on VisDrone2019 and HIT-UAV show consistent gains over the YOLO11n baseline: DPCNet improves mAP@0.5 by 2.0% and 5.1%, respectively, with higher precision and recall, especially for small, dense, low-light, and occluded targets. Despite modest computational overhead from cross-path interactions, the parameter count is reduced by about 45%, indicating a compact and robust solution for small object detection in challenging UAV scenarios.

Schlagwörter

MeSH-Schlagwörter (MEDLINE)

Unmanned Aerial DevicesImage Processing, Computer-AssistedDetection AlgorithmsAlgorithms

Autor:innen und Institutionen

  • School of Intelligent Manufacturing and Electrical Engineering, Guangzhou Institute of Science and Technology, Guangzhou, Guangdong Province, China.

ORCID-Profile: Linfeng Jia, Yafeng Zhu

Förderung

  • Guangdong Provincial Undergraduate Universities Teaching Quality and Teaching Reform Program—Excellence Program for Talent Training in Intelligent Manufacturing
  • Guangdong Provincial Key Discipline Construction Project for Scientific Research Capacity Enhancement
  • University Level First Class Undergraduate Major Program of Guangzhou Institute of Science and Technology—Robotics Engineering
  • the University Level First Class Undergraduate Major Program of Guangzhou Institute of Science and Technology—Robotics Engineering
  • the Guangdong Provincial Key Discipline Construction Project for Scientific Research Capacity Enhancement

Zitieren

APA

Jia, L., Zhu, Y., & Li, B. (2026). DPCNet: A dual path cross perception network for small object detection in UAV imagery. PLOS One, 21(3), e0344091. https://doi.org/10.1371/journal.pone.0344091

Vancouver

Jia L, Zhu Y, Li B. DPCNet: A dual path cross perception network for small object detection in UAV imagery. PLOS One. 2026;21(3):e0344091. doi:10.1371/journal.pone.0344091

BibTeX
@article{Jia2026,
  author = {Jia, Linfeng and Zhu, Yafeng and Li, Bin},
  title = {DPCNet: A dual path cross perception network for small object detection in UAV imagery},
  journal = {PLOS One},
  year = {2026},
  volume = {21},
  number = {3},
  pages = {e0344091},
  doi = {10.1371/journal.pone.0344091}
}
Kostenlos starten →