Deep 2D Convolutional Neural Network with Deconvolution Layer for Hyperspectral Image Classification

  • Chunyan Yu
  • Fang LiEmail author
  • Chein-I Chang
  • Kun Cen
  • Meng Zhao
Conference paper
Part of the Lecture Notes in Electrical Engineering book series (LNEE, volume 516)


Feature extraction and classification technology based on hyperspectral data have been a hot issue. Recently, the convolutional neural network (CNN) has attracted more attention in the field of hyperspectral image classification. To enhance the feature extracted from the hidden layers, in this paper a deconvolution layer is introduced in the deep 2DCNN model. Analyzing the function of convolution and pooling to determine the structure of the convolutional neural network, deconvolution is used to map low-dimensional features into high-dimensional input; the target pixel and its pixels in a certain neighborhood are input into the network as input data. Experiments on two public available hyperspectral data sets show that the deconvolution layer can better generalize features for the hyperspectral image and the proposed 2DCNN classification method can effectively improve the classification accuracy in comparison with other feature extraction methods.


Deep learning Convolutional neural network Hyperspectral image classification 


  1. 1.
    Chen X. Hyperspectral image classification using deep learning method. China University of Geosciences; 2016.Google Scholar
  2. 2.
    Guo K, Li N. Research on classification of architectural style image based on convolution neural network. In: IEEE, information technology and mechatronics engineering conference. IEEE; 2017. p. 1062–6.Google Scholar
  3. 3.
    Guo LL. Research on image classification algorithm based on deep learning. Doctoral dissertation, China University of Mining and Technology; 2016.Google Scholar
  4. 4.
    Lu X, Chen Y, Li X. Hierarchical recurrent neural hashing for image retrieval with hierarchical convolutional features. IEEE Trans Image Process. 2017;99:1.Google Scholar
  5. 5.
    Song M, Chang CI. A theory of recursive orthogonal subspace projection for hyperspectral imaging. IEEE Trans Geosci Remote Sens. 2015;53(6):3055–72.CrossRefGoogle Scholar
  6. 6.
    Wang Q. Classification for hyperspectral remote sensing image based on deep learning. Doctoral dissertation, Huaqiao University; 2016.Google Scholar
  7. 7.
    Wang Y, Lee LC, Xue B et al. A posteriori hyperspectral anomaly detection for unlabeled classification. IEEE Trans Geosci Remote Sens. 2018:1–16.Google Scholar
  8. 8.
    Yu C, Xue B, Song M, et al. Iterative target-constrained interference-minimized classifier for hyperspectral classification. IEEE J Sel Top Appl Earth Obs Remote Sens. 2018;99:1–23.Google Scholar
  9. 9.
    Yu C, Lee LC, Chang CI et al. Band-specified virtual dimensionality for band selection: an orthogonal subspace projection approach. IEEE Trans Geosci Remote Sens. 2018;99:1–11.Google Scholar
  10. 10.
    Zhao M, Zhang J, Tao C. Land use classification in remote sensing images based on deep convolution neural network. In: Asia-Pacific computational intelligence and information technology conference. Shanghai; 2017.Google Scholar

Copyright information

© Springer Nature Singapore Pte Ltd. 2020

Authors and Affiliations

  • Chunyan Yu
    • 1
  • Fang Li
    • 1
    Email author
  • Chein-I Chang
    • 1
    • 2
  • Kun Cen
    • 1
  • Meng Zhao
    • 1
  1. 1.Dalian Maritime UniversityDalianChina
  2. 2.Department of Computer Science and Electrical EngineeringUniversity of Maryland, Baltimore CountyBaltimoreUSA

Personalised recommendations