EUVIMEDEUVIMED entdecken
Open Access · CC BY 4.0

A UAV-LiDAR based paradigm for rapid forest-vegetation health screening using canopy structural indicators

Zhuo Chen, Jincheng Liu, Zheng Gong, Minghui Wang, Zhen Li, Donghai Yu, Huanhuan Zhao

Forest Ecosystems · 2026-11 · 100562

DOI: 10.1016/j.fecs.2026.100562 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Forest Ecosystems, 2026-11. Verfasst von Zhuo Chen und 6 weiteren Autor:innen. Die Arbeit nennt 88 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0). Gefördert von: Shaanxi Province Key Research and Development Projects, National Natural Science Foundation of China.

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Forest health assessment is essential for ecosystem conservation and climate-change adaptation, yet existing approaches often rely on labor-intensive field surveys or complex multi-source remote-sensing workflows. This study developed a rapid structural screening framework for forest health based on unmanned aerial vehicle light detection and ranging (UAV-LiDAR). Ten canopy structural indicators were extracted from 844 plots across seven climatically distinct regions in China, and five key indicators were retained through hierarchical clustering. Two types of forest health indices (FHIs) were then constructed using eight machine-learning models: additive structural FHIs (AS-FHIs) and probability-weighted structural FHIs (PW-FHIs). Results showed that PW-FHIs generally outperformed AS-FHIs. The best-performing extreme gradient boosting-PW-FHIs (XGB-PW-FHIs) achieved an overall accuracy of 0.7736 ± 0.0538, followed by random forest-PW-FHIs (RF-PW-FHIs) with an accuracy of 0.7653 ± 0.0566, while the best AS-FHIs model, AdaBoost-AS-FHIs, achieved an accuracy of 0.6836 ± 0.0429. Regional threshold recalibration substantially improved AS-FHIs accuracy from 0.6836 to 0.7867, whereas PW-FHIs showed higher global stability, with accuracy changing slightly from 0.7737 to 0.7630 after recalibration. FHIs showed strong discrimination between healthy and unhealthy stands, but the identification of sub-healthy stands remained uncertain. The proposed FHIs are not intended as complete diagnostic indicators of forest health, because single-source LiDAR cannot directly capture physiological, biochemical, or belowground processes. Instead, they provide a low-cost, repeatable, and efficient structural proxy for rapidly identifying stands with potential canopy degradation and prioritizing areas for field verification and targeted management.

Förderung

  • Shaanxi Province Key Research and Development Projects
  • National Natural Science Foundation of China

Zitieren

APA

Chen, Z., Liu, J., Gong, Z., Wang, M., Li, Z., Yu, D., & Zhao, H. (2026). A UAV-LiDAR based paradigm for rapid forest-vegetation health screening using canopy structural indicators. Forest Ecosystems, 100562. https://doi.org/10.1016/j.fecs.2026.100562

Vancouver

Chen Z, Liu J, Gong Z, Wang M, Li Z, Yu D, et al. A UAV-LiDAR based paradigm for rapid forest-vegetation health screening using canopy structural indicators. Forest Ecosystems. 2026:100562. doi:10.1016/j.fecs.2026.100562

BibTeX
@article{Chen2026,
  author = {Chen, Zhuo and Liu, Jincheng and Gong, Zheng and Wang, Minghui and Li, Zhen and Yu, Donghai and Zhao, Huanhuan},
  title = {A UAV-LiDAR based paradigm for rapid forest-vegetation health screening using canopy structural indicators},
  journal = {Forest Ecosystems},
  year = {2026},
  pages = {100562},
  doi = {10.1016/j.fecs.2026.100562}
}
Kostenlos starten →