This work proposes a four–stage hybrid framework for hyperspectral image classification that integrates spectral modeling and spatial feature extraction within a unified pipeline and compares the model with recent baselines, including DaViT.
Abstract
Hyperspectral image (HSI) classification remains a challenging problem in remote sensing. This is mainly due to the high dimensionality of spectral data, the limited availability of labeled samples, and the need to jointly model spectral and spatial information. Recent studies have shown that Transformer encoders are effective for spectral modeling. Convolutional architectures are also widely used for spatial feature extraction. However, their integration remains relatively underexplored, particularly under label–scarce conditions. In this work, we propose a four–stage hybrid framework for hyperspectral image classification. The framework integrates spectral modeling and spatial feature extraction within a unified pipeline. First, a multiscale 1D convolutional module performs spectral reduction using parallel kernels of sizes 3, 7, and 11. These kernels capture spectral patterns at multiple resolutions. The module compresses the original spectral bands into a 32–dimensional embedding while preserving important spectral information through a residual skip connection. Second, a pixel–wise Transformer encoder models long–range spectral dependencies in the compressed token sequence. The encoder uses four-head self-attention to capture non–local spectral relationships. Third, a U-Net based spatial extractor captures spatial context within the hyperspectral scene. Hierarchical encoder–decoder skip connections help preserve spatial information and recover fine–grained boundary details. Finally, a dual–stream ensemble classification head aggregates complementary feature descriptors. It combines global average pooling and global max pooling to improve prediction robustness. We evaluate the proposed architecture on two benchmark datasets: Indian Pines and Pavia University. Experiments are conducted under both standard training conditions (80%) and limited–label settings (5%, 10%, and 15%). We also compare the model with recent baselines, including DaViT. On the Indian Pines dataset with 80% training data, the model achieves 99.85% overall accuracy and a weighted F1 score of 99.53%. The Cohen’s
$$\kappa$$
value reaches 0.9945 while the model uses only 7.7 M parameters.
Hyperspectral image classification is an important research topic in the field of remote sensing. Its core objective is to accurately identify different land-cover categories by jointly exploiting the spatial information and spectral characteristics of pixels. However, the complex and diverse spatial structures of grou...
Li-Guo Wang, Teng Wen, Dan-Feng Liu· Remote Sensing· 0 citations
A hyperspectral conditional diffusion for dense (HCD2) classification framework is proposed in this article, which reformulates HSI classification as a conditional label-map denoising process guided by hyperspectral features, enabling generative dense semantic inference.
Hong-Qiao Wang, Hai-Yan Sun, Xuan-Chen Yan et al.· IEEE Transactions on Geoscie...· 0 citations
Hyperspectral image classification (HSIC) aims to discriminate land cover categories accurately by exploiting rich spatial and spectral information. However, existing methods still suffer from insufficient collaborative modeling of spatial and spectral features, as well as inadequate fusion of local and global informat...
The results suggest that decoupled spatial–spectral representation and adaptive multi-scale fusion can improve hyperspectral image classification under the evaluated benchmark settings.
RS-HCFormer is proposed, a hierarchical conditional encoding Transformer for hyperspectral image classification that substantially outperforms state-of-the-art methods in both accuracy and efficiency.
Qiang Li, Hao-Yue Xu, Jia-Ning Zhang et al.· 3rd International Conference...· 0 citations
A new method for HSI classification by integrating spectral band selection with advanced deep learning techniques is presented, and an optimized band selection method alongside data augmentation using crop-and-resize transformations to enhance feature diversity is proposed.
Eman N. Abdelhafez, Ahmed Hagag, Tamer A. Abassy· Scientific Reports· 0 citations
We use cookies to run the site and, with your consent, for analytics and to show ads.
See our Cookie Policy.