EUVIMEDEUVIMED entdecken
Open Access · CC BY 4.0

TinyCapsViT: Ultra-Lightweight Hyperspectral and Multispectral Image Classification for UAV Edge Deployment Using Capsule Vision Transformers

Sagar Dalai, Marco Moreno, Eoin O’Connell, Thomas Newe, Gerard Dooly

Remote Sensing · 07.08.2026 · 18 (16) · 2661

DOI: 10.3390/rs18162661 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Remote Sensing, online erschienen am 07.08.2026. Verfasst von Sagar Dalai und 4 weiteren Autor:innen. Die Arbeit nennt 36 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0). Gefördert von: University of Limerick, BLUEPOINT project, Climate+ Co-Centre.

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Hyperspectral image (HSI) classification is central to environmental monitoring, yet real-time deployment of deep learning models on resource-constrained edge platforms remains challenging due to high spectral dimensionality and computational overhead. In this paper, we propose TinyCapsViT, an ultra-lightweight hybrid architecture that integrates convolutional feature extraction, transformer-based attention, and capsule-inspired representation learning for efficient HSI classification under strict TinyML constraints. The model employs a minimal convolutional stem using pointwise and depthwise separable convolutions to capture local spectral-spatial features, followed by a compact tokenization strategy and learnable positional embeddings. A lightweight self-attention module enables global context modeling with reduced computational complexity, while a capsule-inspired refinement block with squash nonlinearity and residual scaling enhances feature discrimination. The proposed architecture contains only 2781 trainable parameters, representing up to a 30× reduction compared with larger architectures such as ResNet and Vision Transformer, and approximately 11× and 9.5× fewer parameters than the CNN and 3D-CNN baselines, respectively. Extensive experiments across benchmark hyperspectral datasets demonstrate that TinyCapsViT achieves up to 99.61% overall accuracy and maintains competitive classification performance despite its substantially reduced model complexity. Although TinyCapsViT does not consistently achieve the highest classification accuracy compared with larger baseline models, it provides a favourable trade-off between classification performance and computational efficiency. These results demonstrate the potential of TinyCapsViT as a practical solution for real-time hyperspectral analysis on resource-constrained edge platforms and UAV-based environmental monitoring applications.

Förderung

  • University of Limerick
  • BLUEPOINT project
  • Climate+ Co-Centre

Zitieren

APA

Dalai, S., Moreno, M., O’Connell, E., Newe, T., & Dooly, G. (2026). TinyCapsViT: Ultra-Lightweight Hyperspectral and Multispectral Image Classification for UAV Edge Deployment Using Capsule Vision Transformers. Remote Sensing, 18(16), 2661. https://doi.org/10.3390/rs18162661

Vancouver

Dalai S, Moreno M, O’Connell E, Newe T, Dooly G. TinyCapsViT: Ultra-Lightweight Hyperspectral and Multispectral Image Classification for UAV Edge Deployment Using Capsule Vision Transformers. Remote Sensing. 2026;18(16):2661. doi:10.3390/rs18162661

BibTeX
@article{Dalai2026,
  author = {Dalai, Sagar and Moreno, Marco and O’Connell, Eoin and Newe, Thomas and Dooly, Gerard},
  title = {TinyCapsViT: Ultra-Lightweight Hyperspectral and Multispectral Image Classification for UAV Edge Deployment Using Capsule Vision Transformers},
  journal = {Remote Sensing},
  year = {2026},
  volume = {18},
  number = {16},
  pages = {2661},
  doi = {10.3390/rs18162661}
}
Kostenlos starten →