EUVIMEDEUVIMED entdecken
Open Access · CC BY 4.0

Unsupervised Hyperspectral Unmixing Based on Multi-Faceted Graph Representation and Curriculum Learning

Ran Liu, Junfeng Pu, Yanru Chen, Yanling Miao, Dawei Liu, Qi Wang

Remote Sensing · 21.04.2026 · 18 (8) · 1250

DOI: 10.3390/rs18081250 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Remote Sensing, online erschienen am 21.04.2026. Verfasst von Ran Liu und 5 weiteren Autor:innen. Die Arbeit nennt 61 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0).

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Hyperspectral unmixing aims to estimate endmember spectra and their corresponding abundance fractions at the subpixel scale, which is a critical preprocessing step for quantitative analysis of hyperspectral remote sensing imagery. While deep learning-based methods have achieved remarkable progress, three fundamental challenges remain: (i) reliance on a single shared spatial prior that cannot decouple the heterogeneous spatial patterns of different land covers; (ii) the lack of synergy in jointly optimizing endmember extraction and abundance estimation; (iii) the poor robustness of unsupervised training to complex mixtures, noise, and class imbalance. To address these issues, we propose a novel unsupervised unmixing framework that integrates adaptive orthogonal multi-faceted graph representation with curriculum learning. Specifically, we design an Adaptive Orthogonal Multi-Faceted Graph Generator (AOMFG) to learn a set of independent orthogonal graph structures, achieving spatially informed decoupling of land cover patterns. Then, a dual-branch collaborative optimization network is constructed: a Graph Convolutional Network (GCN) branch that incorporates the learned spatial topological priors for abundance estimation, and a 1D Convolutional Neural Network (1DCNN) branch that employs a query-attention mechanism to adaptively aggregate pure spectral features for endmember extraction. Finally, we introduce a three-stage curriculum learning strategy that progressively fine-tunes the model, which significantly enhances its performance. Extensive experiments on three widely used real-world benchmark datasets demonstrate that our proposed framework consistently outperforms state-of-the-art methods in both endmember extraction and abundance estimation accuracy. Comprehensive ablation studies, parameter sensitivity analysis, and noise robustness tests further validate the effectiveness of each core component.

Zitieren

APA

Liu, R., Pu, J., Chen, Y., Miao, Y., Liu, D., & Wang, Q. (2026). Unsupervised Hyperspectral Unmixing Based on Multi-Faceted Graph Representation and Curriculum Learning. Remote Sensing, 18(8), 1250. https://doi.org/10.3390/rs18081250

Vancouver

Liu R, Pu J, Chen Y, Miao Y, Liu D, Wang Q. Unsupervised Hyperspectral Unmixing Based on Multi-Faceted Graph Representation and Curriculum Learning. Remote Sensing. 2026;18(8):1250. doi:10.3390/rs18081250

BibTeX
@article{Liu2026,
  author = {Liu, Ran and Pu, Junfeng and Chen, Yanru and Miao, Yanling and Liu, Dawei and Wang, Qi},
  title = {Unsupervised Hyperspectral Unmixing Based on Multi-Faceted Graph Representation and Curriculum Learning},
  journal = {Remote Sensing},
  year = {2026},
  volume = {18},
  number = {8},
  pages = {1250},
  doi = {10.3390/rs18081250}
}
Kostenlos starten →