Skip to main page content
U.S. flag

An official website of the United States government

Dot gov

The .gov means it’s official.
Federal government websites often end in .gov or .mil. Before sharing sensitive information, make sure you’re on a federal government site.

Https

The site is secure.
The https:// ensures that you are connecting to the official website and that any information you provide is encrypted and transmitted securely.

Access keys NCBI Homepage MyNCBI Homepage Main Content Main Navigation
. 2021 May 10;12(5):545.
doi: 10.3390/mi12050545.

Hybrid Dilated Convolution with Multi-Scale Residual Fusion Network for Hyperspectral Image Classification

Chenming Li  1 , Zelin Qiu  1 , Xueying Cao  1 , Zhonghao Chen  1 , Hongmin Gao  1 , Zaijun Hua  1
Affiliations

Hybrid Dilated Convolution with Multi-Scale Residual Fusion Network for Hyperspectral Image Classification

Chenming Li et al. Micromachines (Basel). .

Abstract

The convolutional neural network (CNN) has been proven to have better performance in hyperspectral image (HSI) classification than traditional methods. Traditional CNN on hyperspectral image classification is used to pay more attention to spectral features and ignore spatial information. In this paper, a new HSI model called local and hybrid dilated convolution fusion network (LDFN) was proposed, which fuses the local information of details and rich spatial features by expanding the perception field. The details of our local and hybrid dilated convolution fusion network methods are as follows. First, many operations are selected, such as standard convolution, average pooling, dropout and batch normalization. Then, fusion operations of local and hybrid dilated convolution are included to extract rich spatial-spectral information. Last, different convolution layers are gathered into residual fusion networks and finally input into the softmax layer to classify. Three widely hyperspectral datasets (i.e., Salinas, Pavia University and Indian Pines) have been used in the experiments, which show that LDFN outperforms state-of-art classifiers.

Keywords: HSI classification; local and hybrid dilated convolution; residual fusion networks.

PubMed Disclaimer

Conflict of interest statement

The authors declare no conflict of interest.

Figures

Figure 1
Figure 1
Standard and dilated convolution.
Figure 2
Figure 2
Hybrid dilated convolution.
Figure 3
Figure 3
The flowchart of the LDFN model.
Figure 4
Figure 4
Indian Pines image. (a) Sample band of Indian Pines dataset. (b) Ground truth data. (c) Color band.
Figure 5
Figure 5
Salinas image. (a) Sample band of Salinas dataset. (b) Ground truth data. (c) Color band.
Figure 6
Figure 6
University of Pavia image. (a) Sample band of Pavia University dataset. (b) Ground truth data. (c) Color band.
Figure 7
Figure 7
Overall accuracy (%) with different hyperparameters on three datasets. (a) patch sizes, (b) principal component numbers.
Figure 8
Figure 8
Classification maps for the Indian Pines dataset. (a) SVM:80.01%. (b) 3D-CNN:94.10%. (c) 3D-CAE:92.04%. (d) D-CNN:97.93%. (e) SSRN:98.09%. (f) LDFN:98.54%.
Figure 9
Figure 9
Classification maps for the Salinas dataset. (a) SVM: 85.97% (b) 3D-CNN: 95.24% (c) 3D-CAE: 96.05% (d) D-CNN: 95.35%. (e) SSRN: 98.38%. (f) LDFN: 99.36%.
Figure 10
Figure 10
Classification maps for the University of Pavia dataset. (a) SVM: 89.18% (b) 3D-CNN: 94.33% (c) 3D-CAE: 95.36% (d) D-CNN: 97.19% (e) SSRN: 98.57%. (f) LDFN: 99.19%.
See this image and copyright information in PMC

References

    1. Zheng X., Yuan Y., Lu X. Dimensionality reduction by spatial–spectral preservation in selected bands. IEEE Trans. Geosci. Remote Sens. 2017;55:5185–5197. doi: 10.1109/TGRS.2017.2703598. - DOI
    1. Zhang L., Zhang L., Tao D., Huang X., Du B. Hyperspectral remote sensing image subpixel target detection based on supervised metric learning. IEEE Trans. Geosci. Remote Sens. 2013;52:4955–4965. doi: 10.1109/TGRS.2013.2286195. - DOI
    1. McManamon P.F. Dual use opportunities for EO sensors-how to afford military sensing; Proceedings of the 15th Annual AESS/IEEE Dayton Section Symposium. Sensing the World: Analog Sensors and Systems Across the Spectrum (Cat. No.98EX178); Fairborn, OH, USA. 14–15 May 1998; pp. 49–52. - DOI
    1. Gevaert C.M., Suomalainen J., Tang J., Kooistra L. Generation of spectral–temporal response surfaces by combining multispectral satellite and hyperspectral uav imagery for precision agriculture applications. IEEE J. Sel. Top. Appl. Earth Obs. Remote Sens. 2015;8:3140–3146. doi: 10.1109/JSTARS.2015.2406339. - DOI
    1. Lu B., Dao P.D., Liu J., He Y., Shang J. Recent advances of hyperspectral imaging technology and applications in agriculture. Remote Sens. 2020;12:2659. doi: 10.3390/rs12162659. - DOI

LinkOut - more resources