Abstract:The utilization of Convolutional Neural Networks (CNNs) in hyperspectral image (HSI) classification has become commonplace. However, traditional CNNs cannot fully extract the features of HSI and are prone to gradient vanishing when the network layer is deepened. We suggest a 2D–3D hybrid convolution and pre-activated residual networks-based HSI classification (HSIC) approach to tackle these problems. Firstly, the joint spatial–spectral features of HSI are extracted by a two-layer 3D convolution. Secondly, combining the advantages of 2D and 3D convolution to construct a spatial–spectral feature extraction module based on pre-activated residual networks, which can accelerate the convergence speed of the model while enhancing the capability of advanced spatial semantic feature extraction of HSI. Then, multiple residual modules are connected to take advantage of the different forms of features extracted by each convolutional layer, while multi-feature fusion is performed between blocks to achieve feature complementarity. Finally, a long-distance residual connection is introduced to fuse the shallow and deep features effectively, which further strengthens the expression ability of features. The results of the experiments conducted on three HSIs show that the overall classification accuracy of the model reaches 99.56%, 99.45% and 99.43%, respectively, when 10%, 1% and 1% of samples are randomly selected for training in each ground object class. Compared with other related CNN-based HSI classification models, our model can obtain higher classification accuracy. Consequently, the suggested method is capable of achieving feature reuse and obtaining deep high-level spatial–spectral features with superior discriminative and robustness, and its classification performance is superior to that of existing state-of-the-art methods.

The real-time and stack fusion enhanced dual-channel network with attention modules for fast hyperspectral image classification

Hyperspectral and Multispectral Image Fusion Based on Deep Attention Network.

A Multiscale Dual-Branch Feature Fusion and Attention Network for Hyperspectral Images Classification

Hyperspectral Image Classification Based on Dual-Scale Dense Network with Efficient Channel Attentional Feature Fusion

HAMNet: hyperspectral image classification based on hybrid neural network with attention mechanism and multi-scale feature fusion

Subpixel Multilevel Scale Feature Learning and Adaptive Attention Constraint Fusion for Hyperspectral Image Classification

A Novel 2D-3D CNN with Spectral-Spatial Multi-Scale Feature Fusion for Hyperspectral Image Classification

An Effective Hyperspectral Image Classification Network Based on Multi-Head Self-Attention and Spectral-Coordinate Attention

Multiscale Feature Fusion Network Incorporating 3D Self-Attention for Hyperspectral Image Classification

Fusing Spatial Attention with Spectral-Channel Attention Mechanism for Hyperspectral Image Classification via Encoder-Decoder Networks

A Cross-Channel Dense Connection and Multi-Scale Dual Aggregated Attention Network for Hyperspectral Image Classification

An Efficient Cross-Modality Self-Calibrated Network for Hyperspectral and Multispectral Image Fusion

Dual-Branch Attention-Assisted CNN for Hyperspectral Image Classification

Multi-layer feature fusion for hyperspectral image classification

Hierarchical Attention and Parallel Filter Fusion Network for Multisource Data Classification

Hybrid 2D–3D convolution and pre-activated residual networks for hyperspectral image classification

Attention Multibranch Convolutional Neural Network for Hyperspectral Image Classification Based on Adaptive Region Search

Hyperspectral Image Classification Based on Two-Branch Spectral–Spatial-Feature Attention Network

Hyperspectral Image Spectral–Spatial Classification Method Based on Deep Adaptive Feature Fusion

A Triple-Path Spectral–Spatial Network With Interleave-Attention for Hyperspectral Image Classification