EUVIMEDEUVIMED entdecken
Crossref

Lightweight Detection of Forest Tree Health Anomalies in Complex UAV Scenes via Dynamic Feature Fusion and Background Suppression

Junxi Wang, Bowei Wang, Yaodan Chi, Yao Zhang, Jing Hu, Shaohui Han, Mengnan Liu, Mingyan Gao

Concurrency and Computation: Practice and Experience · 2026-11 · 38 (22)

DOI: 10.1002/cpe.71030 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Concurrency and Computation: Practice and Experience, online erschienen am 11.11.2026. Verfasst von Junxi Wang und 7 weiteren Autor:innen. Die Arbeit nennt 50 Literaturangaben.

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Unmanned aerial vehicle imagery provides an efficient data source for large‐scale monitoring of forest tree health anomalies. However, detecting unhealthy trees in complex UAV forest scenes remains challenging because target trees may be small, densely distributed, weakly bounded, and visually similar to surrounding vegetation. In addition, background interference from vegetation textures, shadows, bare land, and stones can lead to missed detections, false positives, and localization errors. To address these challenges, this study proposes a YOLO11‐based detection model termed Multi‐scale Attention and Dynamic Convolution YOLO (MADC‐YOLO). Specifically, a statistics‐aware dynamic convolution module (SA‐DyConv) is designed to enhance input‐adaptive feature extraction; a multi‐scale large‐kernel selective attention‐enhanced C3k2 module (MLSA‐C3k2) is employed to integrate contextual information from multiple receptive fields; and a forest background suppression module (FBS) is introduced to filter features in complex backgrounds. On a forest tree health anomaly detection dataset, MADC‐YOLO achieved Precision, Recall, mAP@0.5, and mAP@0.5:0.95 values of 88.36%, 86.39%, 93.31%, and 66.07%, respectively. Ablation experiments further confirmed the contributions of the proposed modules to the overall test‐set metrics. Compared with the other models, MADC‐YOLO achieves higher detection accuracy and greater stability, indicating its potential for applications such as intelligent forestry inspection and early anomaly monitoring.

Zitieren

APA

Wang, J., Wang, B., Chi, Y., Zhang, Y., Hu, J., Han, S., Liu, M., & Gao, M. (2026). Lightweight Detection of Forest Tree Health Anomalies in Complex UAV Scenes via Dynamic Feature Fusion and Background Suppression. Concurrency and Computation: Practice and Experience, 38(22). https://doi.org/10.1002/cpe.71030

Vancouver

Wang J, Wang B, Chi Y, Zhang Y, Hu J, Han S, et al. Lightweight Detection of Forest Tree Health Anomalies in Complex UAV Scenes via Dynamic Feature Fusion and Background Suppression. Concurrency and Computation: Practice and Experience. 2026;38(22). doi:10.1002/cpe.71030

BibTeX
@article{Wang2026,
  author = {Wang, Junxi and Wang, Bowei and Chi, Yaodan and Zhang, Yao and Hu, Jing and Han, Shaohui and Liu, Mengnan and Gao, Mingyan},
  title = {Lightweight Detection of Forest Tree Health Anomalies in Complex UAV Scenes via Dynamic Feature Fusion and Background Suppression},
  journal = {Concurrency and Computation: Practice and Experience},
  year = {2026},
  volume = {38},
  number = {22},
  doi = {10.1002/cpe.71030}
}
Kostenlos starten →