EUVIMEDEUVIMED entdecken
Crossref

CGMSANet: Hyperspectral Image Classification through Channel-Grouped Multi-Scale Feature Fusion and Attention Mechanisms

Mengyan Ye, Xicheng Tan, Chunhui Chen, Yanfei Zhong, Chaopeng Li, Hamza Chaudhary, Ailong Ma, Xiaoliang Meng, Zongyao Sha, Huamin Wang, Jianguang Tu

Photogrammetric Engineering & Remote Sensing · 01.04.2026 · 92 (4) · 309-322

DOI: 10.14358/pers.25-00035r4 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Photogrammetric Engineering & Remote Sensing, erschienen am 01.04.2026. Verfasst von Mengyan Ye und 10 weiteren Autor:innen. Die Arbeit nennt 45 Literaturangaben.

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

To address the limitations of traditional convolutional neural networks in hyperspectral image classification under sample-limited scenarios, including insufficient feature extraction and information redundancy, this paper proposes a novel network model called CGMSANet. This model replaces traditional batch normalization with group normalization (GN) and integrates it with attention mechanisms to effectively mitigate statistical bias in small-sample training, thereby enhancing the model’s generalization capability. Additionally, a dual-path attention module is designed, combining spatial dependency capture (SDC) and spectral feature weighting (SFW). SDC models cross-band global dependencies in the spatial dimension, while SFW dynamically suppresses redundant channels in the spectral dimension, thereby enhancing focus on discriminative features. Experiments on three public data sets—HongHu, LongKou, and HanChuan—demonstrate that CGMSANet achieves superior performance in terms of F1 score and classification accuracy. Specifically, on the HongHu data set, the average F1 score reaches 96.93%, with a 57% reduction in computational complexity and a 2.45× improvement in inference speed compared to the transformer-based SpectralFormer model. Ablation studies further validate the collaborative effectiveness of the GN, SDC, and SFW modules. This model provides an efficient and robust solution for few-shot hyperspectral classification tasks and shows significant potential for practical applications in agricultural monitoring and resource management. The data are available at https://github.com/YMY666yy/CGMASNet.

Zitieren

APA

Ye, M., Tan, X., Chen, C., Zhong, Y., Li, C., Chaudhary, H., Ma, A., Meng, X., Sha, Z., Wang, H., & Tu, J. (2026). CGMSANet: Hyperspectral Image Classification through Channel-Grouped Multi-Scale Feature Fusion and Attention Mechanisms. Photogrammetric Engineering & Remote Sensing, 92(4), 309-322. https://doi.org/10.14358/pers.25-00035r4

Vancouver

Ye M, Tan X, Chen C, Zhong Y, Li C, Chaudhary H, et al. CGMSANet: Hyperspectral Image Classification through Channel-Grouped Multi-Scale Feature Fusion and Attention Mechanisms. Photogrammetric Engineering & Remote Sensing. 2026;92(4):309-322. doi:10.14358/pers.25-00035r4

BibTeX
@article{Ye2026,
  author = {Ye, Mengyan and Tan, Xicheng and Chen, Chunhui and Zhong, Yanfei and Li, Chaopeng and Chaudhary, Hamza and Ma, Ailong and Meng, Xiaoliang and Sha, Zongyao and Wang, Huamin and Tu, Jianguang},
  title = {CGMSANet: Hyperspectral Image Classification through Channel-Grouped Multi-Scale Feature Fusion and Attention Mechanisms},
  journal = {Photogrammetric Engineering & Remote Sensing},
  year = {2026},
  volume = {92},
  number = {4},
  pages = {309--322},
  doi = {10.14358/pers.25-00035r4}
}
Kostenlos starten →