PS2Mamba: A Pyramid-Based Spectral–Spatial Mamba for Hyperspectral Image Classification

When hyperspectral image (HSI) classification encounters high-dimensional spectral channels, the feature utilization rate is often low due to significant redundancy between channels and uneven discriminatory power. Furthermore, different land cover types exhibit notable differences in scale and morphological changes across space. This structural diversity poses significant challenges to spatial feature modeling. To address this problem, this article proposes a novel framework based on the Mamba model-PS2Mamba for HSI classification. This framework integrates three strategies: spectral fine modeling, global perceptual scale adaptation, and multiscale spatial structure modeling. First, this article designs a statistically enhanced normalized band refinement (SENBR) module, which dynamically enhances or suppresses channel features based on channel correlation, variability, and importance, effectively suppressing redundant and noisy bands. Second, a global-aware scale adaptation (GASA) module is proposed. By incorporating a scale scoring network and a multidirectional modeling mechanism, this module enables adaptive perception and directional enhancement modeling of spatial structures at different scales. Finally, a lightweight multiscale spatial structure extraction module, LightPyramid, is constructed. By enriching spatial semantic information through multibranch parallel convolutions, it enhances the model’s ability to represent complex surface structures, while maintaining spatial resolution. Experimental results on four representative hyperspectral datasets, including Pavia University (PU), Salinas (SA), and two unmanned aerial vehicle (UAV)-based datasets (HongHu and HanChuan), demonstrate that PS2Mamba achieves overall accuracies of 98.77%, 99.61%, 96.20%, and 96.01%, respectively. Compared with existing convolutional neural network (CNN)-, graph convolutional network (GCN)-, Transformer-, and Mamba-based models, PS2Mamba achieves up to 12.91% improvement in accuracy, showing superior generalization and robustness, particularly under small-sample conditions.

• Publication year

  2025

• Venue

  IEEE Transactions on Geoscience and Remote Sensing

• Publication date

  Unknown publication date

• Fields of study

  Computer Science, Engineering, Environmental Science

• Identifiers
  DOI 10.1109/TGRS.2025.3631014
• External record

  Open on Semantic Scholar

• Source metadata

  Semantic Scholar

• No claims are published for this paper.

• No concepts are published for this paper.

• SSATNet: Spectral-spatial attention transformer for hyperspectral corn image classification

  2025cited by this paper
• TBi-Mamba: Rethinking Joint Classification of Hyperspectral and LiDAR Data With Bidirectional Mamba

  2025cited by this paper
• Joint Classification of Hyperspectral and LiDAR Data Using Hierarchical Multimodal Feature Aggregation-Based Multihead Axial Attention Transformer

  2025cited by this paper
• MambaHSI: Spatial–Spectral Mamba for Hyperspectral Image Classification

  2025cited by this paper
• Hyperspectral Image Classification Using Groupwise Separable Convolutional Vision Transformer Network

  2024cited by this paper
• A Spatial-Spectral Transformer for Hyperspectral Image Classification Based on Global Dependencies of Multi-Scale Features

  2024cited by this paper
• Channel-Layer-Oriented Lightweight Spectral–Spatial Network for Hyperspectral Image Classification

  2024cited by this paper
• Hyperspectral image classification based on a novel Lush multi-layer feature fusion bias network

  2024cited by this paper
• PSFormer: Pyramid Superpixel Transformer for Hyperspectral Image Classification

  2024cited by this paper
• Joint Classification of Hyperspectral and LiDAR Data Based on Mamba

  2024cited by this paper
• Spectral-Spatial Mamba for Hyperspectral Image Classification

  2024cited by this paper
• Attention Multihop Graph and Multiscale Convolutional Fusion Network for Hyperspectral Image Classification

  2023cited by this paper
• Spectral–Spatial Morphological Attention Transformer for Hyperspectral Image Classification

  2023cited by this paper
• Global–Local 3-D Convolutional Transformer Network for Hyperspectral Image Classification

  2023cited by this paper
• Detection of Ore‐Forming Elements Migrated From Rock and Soil to Vegetation Leaves Using Hyperspectral Data

  2023cited by this paper
• GTFN: GCN and Transformer Fusion Network With Spatial-Spectral Features for Hyperspectral Image Classification

  2023cited by this paper
• FactoFormer: Factorized Hyperspectral Transformers With Self-Supervised Pretraining

  2023cited by this paper
• Cross-Channel Dynamic Spatial–Spectral Fusion Transformer for Hyperspectral Image Classification

  2023cited by this paper
• Hyperspectral Image Classification Using Group-Aware Hierarchical Transformer

  2022cited by this paper
• Spectral–Spatial Feature Tokenization Transformer for Hyperspectral Image Classification

  2022cited by this paper
• Adaptive Cross-Attention-Driven Spatial–Spectral Graph Convolutional Network for Hyperspectral Image Classification

  2022cited by this paper
• A Disjoint Samples-Based 3D-CNN With Active Transfer Learning for Hyperspectral Image Classification

  2022cited by this paper
• SpectralFormer: Rethinking Hyperspectral Image Classification With Transformers

  2021cited by this paper
• Spatial-Spectral Transformer for Hyperspectral Image Classification

  2021cited by this paper
• Trends in Deep Learning for Medical Hyperspectral Image Analysis

  2020cited by this paper
• Hyperspectral Image Classification With Attention-Aided CNNs

  2020cited by this paper
• Multiscale Dynamic Graph Convolutional Network for Hyperspectral Image Classification

  2019cited by this paper
• HybridSN: Exploring 3-D–2-D CNN Feature Hierarchy for Hyperspectral Image Classification

  2019cited by this paper
• An Augmented Linear Mixing Model to Address Spectral Variability for Hyperspectral Unmixing

  2018cited by this paper
• Semisupervised Hyperspectral Image Classification Based on Generative Adversarial Networks

  2018cited by this paper
• A new deep convolutional neural network for fast hyperspectral image classification

  2017cited by this paper
• The properties of the cornea based on hyperspectral imaging: Optical biomedical engineering perspective

  2016cited by this paper
• Generation of Spectral–Temporal Response Surfaces by Combining Multispectral Satellite and Hyperspectral UAV Imagery for Precision Agriculture Applications

  2015cited by this paper
• Investigation of the random forest framework for classification of hyperspectral data

  2005cited by this paper

• No citing papers are available for this paper.