EUVIMEDEUVIMED entdecken
Open Access · CC BY 4.0

LCFF-Net: A lightweight cross-scale feature fusion network for tiny target detection in UAV aerial imagery

Daoze Tang, Shuyun Tang, Zhipeng Fan iD

PLOS ONE · 2024

DOI: 10.1371/journal.pone.0315267 ↗

Kurz gefasst

Fachartikel in der Zeitschrift PLOS ONE, 2024. Verfasst von Daoze Tang und 2 weiteren Autor:innen. Beteiligte Einrichtung: Harbin University of Commerce, Harbin, China. Hauptthemen laut MEDLINE: Algorithms, Image Processing, Computer-Assisted, Unmanned Aerial Devices. Die Arbeit wurde bisher 6-mal zitiert (Crossref) und nennt 51 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0). Gefördert von: Heilongjiang Postdoctoral Fund to pursue scientific research, Collaborative Innovation Achievement Program of Double First-class Disciplines in Heilongjiang Province, National College Students’ Innovation and Entrepreneurship Training Program of China.

AbstractSchlagwörterBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

In the field of UAV aerial image processing, ensuring accurate detection of tiny targets is essential. Current UAV aerial image target detection algorithms face challenges such as low computational demands, high accuracy, and fast detection speeds. To address these issues, we propose an improved, lightweight algorithm: LCFF-Net. First, we propose the LFERELAN module, designed to enhance the extraction of tiny target features and optimize the use of computational resources. Second, a lightweight cross-scale feature pyramid network (LC-FPN) is employed to further enrich feature information, integrate multi-level feature maps, and provide more comprehensive semantic information. Finally, to increase model training speed and achieve greater efficiency, we propose a lightweight, detail-enhanced, shared convolution detection head (LDSCD-Head) to optimize the original detection head. Moreover, we present different scale versions of the LCFF-Net algorithm to suit various deployment environments. Empirical assessments conducted on the VisDrone dataset validate the efficacy of the algorithm proposed. Compared to the baseline-s model, the LCFF-Net-n model outperforms baseline-s by achieving a 2.8% increase in the mAP 50 metric and a 3.9% improvement in the mAP 50–95 metric, while reducing parameters by 89.7%, FLOPs by 50.5%, and computation delay by 24.7%. Thus, LCFF-Net offers high accuracy and fast detection speeds for tiny target detection in UAV aerial images, providing an effective lightweight solution.

Schlagwörter

MeSH-Schlagwörter (MEDLINE)

AlgorithmsImage Processing, Computer-AssistedUnmanned Aerial DevicesNeural Networks, Computer

Autor:innen und Institutionen

  • Harbin University of Commerce, Harbin, China.
  • Heilongjiang Provincial Key Laboratory of Electronic Commerce and Information Processing, Harbin, China.

ORCID-Profile: Zhipeng Fan

Förderung

  • Heilongjiang Postdoctoral Fund to pursue scientific research
  • Collaborative Innovation Achievement Program of Double First-class Disciplines in Heilongjiang Province
  • National College Students’ Innovation and Entrepreneurship Training Program of China
  • Heilongjiang Provincial College Students’ Innovation and Entrepreneurship Training Programme

Zitieren

APA

Tang, D., Tang, S., & Fan, Z. (2024). LCFF-Net: A lightweight cross-scale feature fusion network for tiny target detection in UAV aerial imagery. PLOS ONE. https://doi.org/10.1371/journal.pone.0315267

Vancouver

Tang D, Tang S, Fan Z. LCFF-Net: A lightweight cross-scale feature fusion network for tiny target detection in UAV aerial imagery. PLOS ONE. 2024. doi:10.1371/journal.pone.0315267

BibTeX
@article{Tang2024,
  author = {Tang, Daoze and Tang, Shuyun and Fan, Zhipeng},
  title = {LCFF-Net: A lightweight cross-scale feature fusion network for tiny target detection in UAV aerial imagery},
  journal = {PLOS ONE},
  year = {2024},
  doi = {10.1371/journal.pone.0315267}
}
Kostenlos starten →