GC ‐ YOLO : An Improved Lightweight Aerial Image Detection Algorithm Based on YOLOv11n
Concurrency and Computation: Practice and Experience · 2026-07 · 38 (14)
DOI: 10.1002/cpe.70876 ↗Kurz gefasst
Fachartikel in der Zeitschrift Concurrency and Computation: Practice and Experience, online erschienen am 19.07.2026. Verfasst von Ting Liu und 1 weiteren Autor:innen. Die Arbeit nennt 18 Literaturangaben.
Abstract (Originalsprache)
Aerial image object detection is challenged by small targets, large scale variation, complex backgrounds, and lightweight deployment requirements. To address these issues, this paper proposes GC‐YOLO, a lightweight detector based on YOLOv11n. GC‐YOLO introduces GSConv to improve feature representation while reducing model complexity. It also adapts the PKINet‐derived Context Anchor Attention (CAA) module into the Neck to strengthen long‐range contextual modeling for small and occluded targets. Finally, EIoU loss is used to improve bounding‐box regression. Experiments on a self‐built aerial image dataset show that GC‐YOLO achieves 68.1% mAP@50 and 48.3% mAP@50:95. Compared with YOLOv11n, GC‐YOLO improves mAP@50 by 6.0 percentage points, reduces parameters from 11.4 M to 10.4 M, and reduces model size from 20.5 to 17.1 MB. Comparisons with classical detectors further show that GC‐YOLO obtains the best overall mAP@50 and mAP@50:95, indicating a favorable balance between detection accuracy and model compactness.
Zitieren
APA
Liu, T., & Zhu, J. (2026). GC ‐ YOLO : An Improved Lightweight Aerial Image Detection Algorithm Based on YOLOv11n. Concurrency and Computation: Practice and Experience, 38(14). https://doi.org/10.1002/cpe.70876
Vancouver
Liu T, Zhu J. GC ‐ YOLO : An Improved Lightweight Aerial Image Detection Algorithm Based on YOLOv11n. Concurrency and Computation: Practice and Experience. 2026;38(14). doi:10.1002/cpe.70876
BibTeX
@article{Liu2026,
author = {Liu, Ting and Zhu, Jiankang},
title = {GC ‐ YOLO : An Improved Lightweight Aerial Image Detection Algorithm Based on YOLOv11n},
journal = {Concurrency and Computation: Practice and Experience},
year = {2026},
volume = {38},
number = {14},
doi = {10.1002/cpe.70876}
}
EUVIMED