C2FNet: Cross-Probabilistic Weak Supervision Learning for High-Resolution Land Cover Enhancement
文献类型：期刊论文
| 作者 | Mwubahimana, Boaz9; Yan, Jianguo7,8; Miao, Dingruibo9; Li, Zhuohong6; Guo, Haonan9; Ma, Le9; Mugabowindekwe, Maurice5; Roy, Swalpa Kumar4; Huang, Xiao3; Nyandwi, Elias10 |
| 刊名 | IEEE TRANSACTIONS ON GEOSCIENCE AND REMOTE SENSING
 |
| 出版日期 | 2025 |
| 卷号 | 63页码:30 |
| 关键词 | Coarse-to-fine networks (C2FNets) cross-resolution learning deep neural networks deep neural networks Earth observation Earth observation land cover mapping land cover mapping probabilistic supervision probabilistic supervision remote sensing remote sensing weakly supervised learning (WSL) weakly supervised learning (WSL) weakly supervised learning (WSL) |
| ISSN号 | 0196-2892 |
| DOI | 10.1109/TGRS.2025.3598681 |
| 产权排序 | 1;2;3;4;5;6;7;8 |
| 英文摘要 | Autonomous large-scale high-resolution land cover (HRLC) mapping remains a major challenge in remote sensing due to the scarcity of reliable training data and resolution mismatches between available labels and input of massive of emerging imagery. Existing global land cover products often suffer from coarse spatial resolution and label noise, limiting their utility for fine-scale urban analysis and environmental monitoring. This article presents C2FNet, a novel Coarse-to-Fine Network designed to generate HRLC maps from noisy, coarse-resolution labels using a weak supervision strategy of the cross-probability. The C2FNet consists of three key modules: 1) edge resolution refinement backbones (ERRBs), which preserve spatial detail via multiscale feature extraction through parallel convolutional branches; 2) unsupervised dynamic shuffle and diagonal annotation (UDSDA), which enhances training reliability by identifying confident regions through spatial-consistency analysis and confidence estimation; and 3) a contrasting self-supervised loss (C2F-Loss) that integrates cross-entropy and cosine similarity terms to mitigate supervision noise and resolution gaps. Evaluations of three benchmark datasets that encompass diverse urban and rural landscapes show that C2FNet achieves state-of-the-art (SoA) performance, with 80.01% overall accuracy (OA) and a Cohen's kappa score of 0.7567, outperforming SoA models with weak supervision. The dataset and code are available at http://drive.google.com/file/d/1X_Fz7LQIeix3rV3K29FBfKiU1WMdROe-/view |
| WOS关键词 | SEMANTIC SEGMENTATION ; CLASSIFICATION |
| WOS研究方向 | Geochemistry & Geophysics ; Engineering ; Remote Sensing ; Imaging Science & Photographic Technology |
| 语种 | 英语 |
| WOS记录号 | WOS:001635756400001 |
| 出版者 | IEEE-INST ELECTRICAL ELECTRONICS ENGINEERS INC |
| 源URL | [http://ir.xao.ac.cn/handle/45760611-7/8410]  |
| 专题 | 研究单元未命名 |
| 通讯作者 | Mwubahimana, Boaz; Yan, Jianguo |
| 作者单位 | 1.African Ctr Excellence Data Sci, Kigali, Rwanda 2.Univ Rwanda, AIMS Res & Innovat Ctr, Kigali, Rwanda 3.Emory Univ, Dept Environm Sci, Atlanta, GA 30322 USA 4.Tezpur Univ, Dept Comp Sci & Engn, Tezpur 784028, Assam, India 5.Univ Copenhagen, Dept Geosci & Nat Resource Management, DK-1350 Copenhagen, Denmark 6.Duke Univ, Nicholas Sch Environm, Durham, NC 27708 USA 7.Chinese Acad Sci, Xinjiang Astron Observ, Urumqi 830011, Peoples R China 8.Wuhan Univ, LIESMARS, Wuhan 430079, Peoples R China 9.Wuhan Univ, State Key Lab Informat Engn Surveying Mapping & Re, Wuhan 430079, Peoples R China 10.Univ Rwanda, Coll Sci & Technol, Ctr Geog Informat Syst & Remote Sensing CGIS, Kigali, Rwanda |
| 推荐引用方式 GB/T 7714 | Mwubahimana, Boaz,Yan, Jianguo,Miao, Dingruibo,et al. C2FNet: Cross-Probabilistic Weak Supervision Learning for High-Resolution Land Cover Enhancement[J]. IEEE TRANSACTIONS ON GEOSCIENCE AND REMOTE SENSING,2025,63:30. |
| APA | Mwubahimana, Boaz.,Yan, Jianguo.,Miao, Dingruibo.,Li, Zhuohong.,Guo, Haonan.,...&Rwanyiziri, Gaspard.(2025).C2FNet: Cross-Probabilistic Weak Supervision Learning for High-Resolution Land Cover Enhancement.IEEE TRANSACTIONS ON GEOSCIENCE AND REMOTE SENSING,63,30. |
| MLA | Mwubahimana, Boaz,et al."C2FNet: Cross-Probabilistic Weak Supervision Learning for High-Resolution Land Cover Enhancement".IEEE TRANSACTIONS ON GEOSCIENCE AND REMOTE SENSING 63(2025):30. |
入库方式： OAI收割
来源：新疆天文台
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。