img

Wechat

Adv search

Journal of Desert Research ›› 2026, Vol. 46 ›› Issue (4): 385-397.DOI: 10.7522/j.issn.1000-694X.2026.00057

Previous Articles     Next Articles

Vegetation extraction methods for desertified areas using unmanned aerial vehicle RGB imagery

Bing Ma1,2(), Guojing Yang1, Guosheng Zan3(), Tao Sun3, Yuanxin Jia3, Lihua Zhou2,4   

  1. 1.State Key Laboratory of Ecological Safety and Sustainable Development in Arid Lands,Northwest Institute of Eco-Environment and Resources,Chinese Academy of Sciences,Lanzhou 730000,China
    2.University of Chinese Academy of Sciences,Beijing 100049,China
    3.Academy of Forestry Inventory and Planning,National Forestry and Grassland Administration,Beijing 100714,China
    4.Institutes of Science and Development,Chinese Academy of Sciences,Beijing 100190,China
  • Received:2025-12-30 Revised:2026-04-22 Online:2026-07-20 Published:2026-08-27
  • Contact: Guosheng Zan

Abstract:

Unmanned aerial vehicle (UAV) remote sensing technology, owing to its high spatial resolution and flexibility, has become an important tool for vegetation monitoring in ecosystems. Existing methods mainly focus on green vegetation (GV); however, in desert ecosystems, non-photosynthetic vegetation (NPV, such as dry branches and small shrubs) exhibits spectral characteristics similar to those of green vegetation and bare soil, making it difficult for current approaches to efficiently distinguish vegetation information. To address this issue, this study conducted UAV aerial surveys in two typical desert sample sites in Hangjin Banner, Inner Mongolia. The spectral characteristics of GV and NPV in the visible bands were analyzed in depth. A joint threshold segmentation method, an automatic threshold segmentation method, and a random forest method were applied to classify and extract green vegetation, non-photosynthetic vegetation, and bare land, respectively, and the classification accuracy and applicability of the three methods were evaluated. The experimental results show that the joint threshold method achieved the highest average accuracy of 0.87, followed by the random forest method with an accuracy of 0.82, while the automatic threshold segmentation method performed the worst, with an accuracy of 0.75. The joint threshold method is easy to operate and computationally efficient, making it suitable for application at the grassroots level. It provides more reliable technical support for monitoring of vegetation, evaluation of ecological restoration, and desertification control in desert ecosystems.

Key words: desertified areas, unmanned aerial vehicle, visible (RGB) bands, vegetation extraction

CLC Number: