EUVIMEDEUVIMED entdecken
Open Access · CC BY 4.0

RSWNet: Lightweight Scale–Spatial–Frequency Feature Learning for UAV-Based Greenhouse Rose Monitoring

Gang Dong, Yingchao Wang, Yulun Chen, Kun Yang, Feng Xue, Wei Wang

Drones · 28.10.2026 · 10 (11) · 808

DOI: 10.3390/drones10110808 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Drones, online erschienen am 28.10.2026. Verfasst von Gang Dong und 5 weiteren Autor:innen. Die Arbeit nennt 78 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0). Gefördert von: Japan Science and Technology Agency, Science and Technology Research Program of the Chongqing Municipal Education Commission, Youth Fund of the Southern Xinjiang Joint Fund of the Natural Science Foundation of Xinjiang Uygur Autonomous Region.

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

UAV-based greenhouse rose monitoring requires accurate detection of rose targets under substantial scale variation, foliage occlusion, background similarity, illumination changes, and motion-induced degradation while maintaining lightweight inference for practical real-time edge deployment. This study proposes RoseScaleWaveNet (RSWNet), a YOLO26n-based lightweight detector that coordinates scale, spatial context, and frequency information through three complementary modules: C3MKIR for multi-receptive-field feature extraction, SEASCPP for position-aware multi-scale contextual modeling, and MWAdaLF for frequency-decoupled cross-level feature fusion. Experiments on a multi-source greenhouse rose dataset annotated with two petal-opening classes show that RSWNet achieves 94.13% precision, 91.95% recall, 95.51% mAP@50, and 59.22% mAP@50:95 with only 2.30 M parameters and an inference speed of 185 FPS. Compared with YOLO26n, mAP@50 and mAP@50:95 are improved by 5.50 and 4.70 percentage points, respectively, while model complexity is slightly reduced. The proposed modules also yield an average mAP@50 gain of 4.24 percentage points when transferred to four representative YOLO architectures. Cross-domain evaluation on six external flower-related datasets further demonstrates consistent improvements under both zero-shot evaluation and dataset-specific retraining. Controlled tests under blur, noise, illumination variation, motion blur, and scale changes characterize robustness to the evaluated image perturbations. Separate benchmarks on desktop and laptop GPUs, CPU-only hardware, and Jetson Orin Nano quantify inference efficiency. These results support the potential of RSWNet for UAV-assisted greenhouse rose detection within the evaluated conditions.

Förderung

  • Japan Science and Technology Agency
  • Science and Technology Research Program of the Chongqing Municipal Education Commission
  • Youth Fund of the Southern Xinjiang Joint Fund of the Natural Science Foundation of Xinjiang Uygur Autonomous Region

Zitieren

APA

Dong, G., Wang, Y., Chen, Y., Yang, K., Xue, F., & Wang, W. (2026). RSWNet: Lightweight Scale–Spatial–Frequency Feature Learning for UAV-Based Greenhouse Rose Monitoring. Drones, 10(11), 808. https://doi.org/10.3390/drones10110808

Vancouver

Dong G, Wang Y, Chen Y, Yang K, Xue F, Wang W. RSWNet: Lightweight Scale–Spatial–Frequency Feature Learning for UAV-Based Greenhouse Rose Monitoring. Drones. 2026;10(11):808. doi:10.3390/drones10110808

BibTeX
@article{Dong2026,
  author = {Dong, Gang and Wang, Yingchao and Chen, Yulun and Yang, Kun and Xue, Feng and Wang, Wei},
  title = {RSWNet: Lightweight Scale–Spatial–Frequency Feature Learning for UAV-Based Greenhouse Rose Monitoring},
  journal = {Drones},
  year = {2026},
  volume = {10},
  number = {11},
  pages = {808},
  doi = {10.3390/drones10110808}
}
Kostenlos starten →