Spatial-Spectral Attention Graph U-Nets for Hyperspectral Image Classification

Graph neural networks (GNNs) have outstanding performance in modeling global spatial information dependence, which makes them very suitable for the complex and diverse distribution of objects in hyperspectral images (HSIs). In recent years, they have been widely used in research on HSI classificatio...

Full description

Saved in:
Bibliographic Details
Published in:IEEE transactions on geoscience and remote sensing 2023, Vol.61, p.1-17
Main Authors: Qu, Kewen, Wang, Chenyang, Li, Zhenqing, Luo, Fangzhou
Format: Article
Language:English
Subjects:
Aggregation
Artificial neural networks
Attention mechanism
Classification
Convolution
Convolutional neural networks
Feature extraction
graph attention network (GAT)
Graph neural networks
graph U-Nets
Graphical representations
hyperspectral image (HSI) classification
Hyperspectral imaging
Image classification
Image segmentation
Kernel
Modules
Neural networks
Pixels
Receptive field
Redundancy
Resource consumption
Spatial data
Spatial dependencies
Citations: Items that this one cites
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:Graph neural networks (GNNs) have outstanding performance in modeling global spatial information dependence, which makes them very suitable for the complex and diverse distribution of objects in hyperspectral images (HSIs). In recent years, they have been widely used in research on HSI classification. However, the high dimensionality and high redundancy between bands of HSIs make it difficult to explore deep spectral features, resulting in high computational and storage costs for GNNs. In addition, traditional convolutional neural networks (CNNs) preprocess images to obtain pixel-level features, but the local receptive field of their convolutional kernels cannot model nonlocal spatial dependencies, leading to extracted features that lack diversity. To address these issues, this article proposes a graph convolutional network that integrates spatial and spectral attention using the graph U-Nets architecture for HSI classification. First, a spectral pixel sequence (SPS) extraction module is constructed in the encoder stage based on spectral attention to reduce the influence of redundant bands and enable graph convolution to extract more discriminative spectral feature representations. Simultaneously, a lightweight graph attention network (LwGAT) is designed to reduce computing resource consumption during similarity matrix calculation. Second, the multiscale feature aggregation module (MFAM) based on spatial attention is used to extract pixel-level features from the original hyperspectral data, which provides a more diverse spatial-spectral feature for the graph U-Nets by fusing spatial contextual information at different scales. Finally, experiments on four widely used HSI datasets demonstrate that the proposed method achieves better classification performance than other advanced classification methods.
ISSN:0196-2892
1558-0644
DOI:10.1109/TGRS.2023.3324977