Open Access iconOpen Access

ARTICLE

crossmark

Deep 3D-Multiscale DenseNet for Hyperspectral Image Classification Based on Spatial-Spectral Information

Haifeng Song1, Weiwei Yang1,*, Haiyan Yuan2, Harold Bufford3

1 School of Electronics and Information Engineering (School of Big Data Science), Taizhou University, Taizhou, 318000, China
2 College of Science, Heilongjiang Institute of Technology, Harbin, 150050, China
3 Departments of Interactive Technology, Animax Designs, Nashville, 37207, USA

* Corresponding Author: Weiwei Yang. Email: email

Intelligent Automation & Soft Computing 2020, 26(6), 1441-1458. https://doi.org/10.32604/iasc.2020.011988

Abstract

There are two main problems that lead to unsatisfactory classification performance for hyperspectral remote sensing images (HSIs). One issue is that the HSI data used for training in deep learning is insufficient, therefore a deeper network is unfavorable for spatial-spectral feature extraction. The other problem is that as the depth of a deep neural network increases, the network becomes more prone to overfitting. To address these problems, a dual-channel 3D-Multiscale DenseNet (3DMSS) is proposed to boost the discriminative capability for HSI classification. The proposed model has several distinct advantages. First, the model consists of dual channels that can extract both spectral and spatial features, both of which are used in HSI classification. Therefore, the classification accuracy can be improved. Second, the 3D-Multiscale DenseNet is used to extract the spectral and spatial features which make full use of the HSI cube. The discriminant features for image classification are extracted and the spectral and spatial features are fused, which can alleviate the problem of low accuracy caused by limited training samples. Third, the connections between different layers are established using a residual dense block, and the feature maps of each layer are fully utilized to further alleviate the vanishing gradient problem. Qualitative classification experiments are reported that show the effectiveness of the proposed method. Compared with existing HSI classification techniques, the proposed method is highly suitable for HSI classification, especially for datasets with fewer training samples. The best overall accuracy of 99.36%, 99.86%, and 99.99% were obtained for the Indian Pines, KSC, and SA datasets, which showed an effective improvement of the classification accuracy.

Keywords


Cite This Article

APA Style
Song, H., Yang, W., Yuan, H., Bufford, H. (2020). Deep 3d-multiscale densenet for hyperspectral image classification based on spatial-spectral information. Intelligent Automation & Soft Computing, 26(6), 1441-1458. https://doi.org/10.32604/iasc.2020.011988
Vancouver Style
Song H, Yang W, Yuan H, Bufford H. Deep 3d-multiscale densenet for hyperspectral image classification based on spatial-spectral information. Intell Automat Soft Comput . 2020;26(6):1441-1458 https://doi.org/10.32604/iasc.2020.011988
IEEE Style
H. Song, W. Yang, H. Yuan, and H. Bufford "Deep 3D-Multiscale DenseNet for Hyperspectral Image Classification Based on Spatial-Spectral Information," Intell. Automat. Soft Comput. , vol. 26, no. 6, pp. 1441-1458. 2020. https://doi.org/10.32604/iasc.2020.011988

Citations




cc Copyright © 2020 The Author(s). Published by Tech Science Press.
This work is licensed under a Creative Commons Attribution 4.0 International License , which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
  • 2242

    View

  • 1362

    Download

  • 2

    Like

Share Link