Pyramid Cascaded Convolutional Neural Network with Graph Convolution for Hyperspectral Image Classification
-
Published:2024-08-11
Issue:16
Volume:16
Page:2942
-
ISSN:2072-4292
-
Container-title:Remote Sensing
-
language:en
-
Short-container-title:Remote Sensing
Author:
Pan Haizhu12, Yan Hui1, Ge Haimiao12ORCID, Wang Liguo3, Shi Cuiping4ORCID
Affiliation:
1. College of Computer and Control Engineering, Qiqihar University, Qiqihar 161000, China 2. Heilongjiang Key Laboratory of Big Data Network Security Detection and Analysis, Qiqihar University, Qiqihar 161000, China 3. College of Information and Communication Engineering, Dalian Nationalities University, Dalian 116000, China 4. College of Telecommunication and Electronic Engineering, Qiqihar University, Qiqihar 161000, China
Abstract
Convolutional neural networks (CNNs) and graph convolutional networks (GCNs) have made considerable advances in hyperspectral image (HSI) classification. However, most CNN-based methods learn features at a single-scale in HSI data, which may be insufficient for multi-scale feature extraction in complex data scenes. To learn the relations among samples in non-grid data, GCNs are employed and combined with CNNs to process HSIs. Nevertheless, most methods based on CNN-GCN may overlook the integration of pixel-wise spectral signatures. In this paper, we propose a pyramid cascaded convolutional neural network with graph convolution (PCCGC) for hyperspectral image classification. It mainly comprises CNN-based and GCN-based subnetworks. Specifically, in the CNN-based subnetwork, a pyramid residual cascaded module and a pyramid convolution cascaded module are employed to extract multiscale spectral and spatial features separately, which can enhance the robustness of the proposed model. Furthermore, an adaptive feature-weighted fusion strategy is utilized to adaptively fuse multiscale spectral and spatial features. In the GCN-based subnetwork, a band selection network (BSNet) is used to learn the spectral signatures in the HSI using nonlinear inter-band dependencies. Then, the spectral-enhanced GCN module is utilized to extract and enhance the important features in the spectral matrix. Subsequently, a mutual-cooperative attention mechanism is constructed to align the spectral signatures between BSNet-based matrix with the spectral-enhanced GCN-based matrix for spectral signature integration. Abundant experiments performed on four widely used real HSI datasets show that our model achieves higher classification accuracy than the fourteen other comparative methods, which shows the superior classification performance of PCCGC over the state-of-the-art methods.
Reference69 articles.
1. Lorenz, S., Salehi, S., Kirsch, M., Zimmermann, R., Unger, G., Vest Sørensen, E., and Gloaguen, R. (2018). Radiometric Correction and 3D Integration of Long-Range Ground-Based Hyperspectral Imagery for Mineral Exploration of Vertical Outcrops. Remote Sens., 10. 2. Rajabi, R., Zehtabian, A., Singh, K.D., Tabatabaeenejad, A., Ghamisi, P., and Homayouni, S. (2024). Editorial: Hyperspectral Imaging in Environmental Monitoring and Analysis. Front. Environ. Sci., 11. 3. Ground-Based Hyperspectral Image Surveillance System for Explosive Detection: Methods, Experiments, and Comparisons;Geneci;IEEE J. Sel. Top. Appl. Earth Obs. Remote Sens.,2023 4. Waczak, J., Aker, A., Wijeratne, L.O.H., Talebi, S., Fernando, A., Dewage, P.M.H., Iqbal, M., Lary, M., Schaefer, D., and Lary, D.J. (2024). Characterizing Water Composition with an Autonomous Robotic Team Employing Comprehensive In Situ Sensing, Hyperspectral Imaging, Machine Learning, and Conformal Prediction. Remote Sens., 16. 5. A Collaborative Approach for the Identification of Thermal Hot-Spots: From Remote Sensing Data to Urban Planning Interventions;Gallacher;AGILE GIScience Ser.,2024
|
|