Feature-model strategies for remapping 1:1,000,000 vegetation map: application to the Tibetan Plateau
文献类型：期刊论文
| 作者 | Wu, Fan2,3; Zhou, Guangsheng1; Ren, Hongrui2,3 |
| 刊名 | ADVANCES IN SPACE RESEARCH
 |
| 出版日期 | 2026-05-01 |
| 卷号 | 77期号:9页码:8982-8996 |
| 关键词 | Vegetation map Remap Machine learning Feature selection |
| ISSN号 | 0273-1177 |
| DOI | 10.1016/j.asr.2026.02.087 |
| 产权排序 | 2 |
| 文献子类 | Article |
| 英文摘要 | Obtaining high-precision vegetation maps is essential for ecological monitoring and natural resource management. However, the existing 1:1,000,000 vegetation maps are manually delineated, which introduces uncertainty and subjectivity in defining boundaries, limiting their utility for detailed management and research. This study demonstrates a reproducible method for reconstructing historical vegetation boundaries with enhanced spatial precision, with a case study on the Tibetan Plateau (TP). Landsat imagery was used as the primary data source, and recursive feature elimination (RFE) was applied to optimize feature sets. Based on selected features and three machine learning models: support vector machine (SVM), gradient boosting tree (GBT), and random forest (RF), 15 classification strategies were designed to evaluate the impact of different feature-model combinations. The results show that: (1) the strategy combining the RF classifier with remote sensing, topography, and climate features optimized through RFE achieved the highest accuracy; (2) the overall accuracy of the remapped vegetation map based on this optimal strategy is 89.03%, with a Kappa coefficient of 0.88; (3) the vegetation map produced in this study has high accuracy and improves delineation accuracy compared with existing maps, showing strong consistency with actual land surface conditions as observed in remote sensing imagery. These findings provide a reliable technical approach for remapping large-scale vegetation maps, and the final vegetation map also offers valuable basic data for vegetation change research on the TP. (c) 2026 COSPAR. Published by Elsevier B.V. All rights are reserved, including those for text and data mining, AI training, and similar technologies. |
| URL标识 | 查看原文 |
| WOS关键词 | DIFFERENCE WATER INDEX ; LEAF-AREA INDEX ; CLIMATE-CHANGE ; LANDSAT TM ; CHINA ; CLASSIFICATION ; DATASET ; IMPACTS ; NDWI |
| WOS研究方向 | Engineering ; Astronomy & Astrophysics ; Geology ; Meteorology & Atmospheric Sciences |
| 语种 | 英语 |
| WOS记录号 | WOS:001746298300001 |
| 出版者 | ELSEVIER SCI LTD |
| 源URL | [http://ir.igsnrr.ac.cn/handle/311030/221540]  |
| 专题 | 资源与环境信息系统国家重点实验室_外文论文 |
| 通讯作者 | Zhou, Guangsheng; Ren, Hongrui |
| 作者单位 | 1.Chinese Acad Meteorol Sci, State Key Lab Severe Weather, Beijing 100081, Peoples R China 2.Chinese Acad Sci, Inst Geog Sci & Nat Resources Res, State Key Lab Resources & Environm Informat Syst, Beijing, Peoples R China; 3.Taiyuan Univ Technol, Dept Geomatics, Taiyuan 030024, Peoples R China; |
| 推荐引用方式 GB/T 7714 | Wu, Fan,Zhou, Guangsheng,Ren, Hongrui. Feature-model strategies for remapping 1:1,000,000 vegetation map: application to the Tibetan Plateau[J]. ADVANCES IN SPACE RESEARCH,2026,77(9):8982-8996. |
| APA | Wu, Fan,Zhou, Guangsheng,&Ren, Hongrui.(2026).Feature-model strategies for remapping 1:1,000,000 vegetation map: application to the Tibetan Plateau.ADVANCES IN SPACE RESEARCH,77(9),8982-8996. |
| MLA | Wu, Fan,et al."Feature-model strategies for remapping 1:1,000,000 vegetation map: application to the Tibetan Plateau".ADVANCES IN SPACE RESEARCH 77.9(2026):8982-8996. |
入库方式： OAI收割
来源：地理科学与资源研究所
浏览0
下载0
收藏0
其他版本
除非特别说明，本系统中所有内容都受版权保护，并保留所有权利。