Spectral Tensor Synthesis Analysis for Hyperspectral Image Spectral–Spatial Feature Extraction

  • Ronghua YanEmail author
  • Jinye Peng
  • Dongmei Ma
  • Desheng Wen
Research Article


Feature extraction is a preprocessing step for hyperspectral image classification. Principal component analysis only uses the spectral information, but it does not use spatial information of a hyperspectral image. Both spatial and spectral information are used when hyperspectral image is modelled as tensor, that is, decreasing the noise on spatial dimension and reducing the dimension on a spectral dimension at the same time. However, in this model, a hyperspectral image is modelled only as a data cube. The factors affecting the spectral features of ground objects is not considered and these factors are barely distinguished. This means that further improving classification is very difficult. Therefore, a new model on hyperspectral image is proposed by the authors. In the new model, many factors that impact the spectral features of ground objects are synthesized as the within-class factor. The within-class factor, the class factor and the pixel spectral are selected as a mode, respectively. The pixel spectrals in the training set are modelled as a third-order tensor. The experiment results indicate that the new method improves the classification compared with the previous methods.


Feature extraction Tensor processing Hyperspectral image Spectral tensor 



This work is supported by the National Natural Science Foundation (Grant No. 61272285) and Program for Changjiang Scholars and Innovative Research Team in University (IRT13090).


  1. Akbari, D. (2017). Improving spectral–spatial classification of hyperspectral imagery using spectral dimensionality reduction based on weighted genetic algorithm. Journal of the Indian Society of Remote Sensing, 45(6), 927–937.CrossRefGoogle Scholar
  2. Bourennane, S., & Fossati, C. (2015). Dimensionality reduction and coloured noise removal from hyperspectral images. Remote Sensing Letters, 6, 854–863.CrossRefGoogle Scholar
  3. Gao, Y., Wang, X., Cheng, Y., & Wang, Z. (2015). Dimensionality reduction for hyperspectral data based on class-aware tensor neighborhood graph and patch alignment. IEEE Transactions on Neural Networks, 26, 1582–1593.CrossRefGoogle Scholar
  4. Landgrebe, D. (2003). Signal theory methods in multispectral remote sensing. New Jersey: Wiley.CrossRefGoogle Scholar
  5. Lathauwer, D., Moor, B. D., & Vandewalle, J. (2000). On the best rank-(r1, rn) approximation of higher-order tensors. SIAM Journal of Matrix Analysis and Applications, 21, 1324–1342.CrossRefGoogle Scholar
  6. Letexier, D., Bourennane, S., & Blanc-Talon, J. (2008). Nonorthogonal tensor matricization for hyperspectral image filtering. IEEE Geoscience and Remote Sensing Letters, 5(1), 3–7.CrossRefGoogle Scholar
  7. Li, C., Ma, Y., Huang, J., Mei, X., & Ma, J. (2015). Hyperspectral image denoising using the robust low-rank tensor recovery. Journal of the Optical Society of America A: Optics, Image Science, and Vision, 32, 1604–1612.CrossRefGoogle Scholar
  8. Ma, L., et al. (2016). Graph-based semi-supervised learning for spectral-spatial hyperspectral image classification. Pattern Recognition Letters, 83, 133–142.CrossRefGoogle Scholar
  9. Michael, D. F., & Mersereau, R. M. (2005). On the impact of PCA dimension reduction for hyperspectral detection of difficult targets. IEEE Geoscience and Remote Sensing Letters, 2, 192–195.CrossRefGoogle Scholar
  10. Muti, D., & Bourennane, S. (2003). Multidimensional signal processing using lower-rank tensor approximation. IEEE ICASSP, 3, 457–460.Google Scholar
  11. Muti, D., & Bourennane, S. (2005). Multidimensional filtering based on a tensor approach. Signal Process, 85, 2338–2353.CrossRefGoogle Scholar
  12. Pu, R., & Gong, P. (2010). Hyperspectral remote sensing and its applications. Beijing, China: Higher Education Press.Google Scholar
  13. Qiao, X., Yuan, D., & Li, H. (2017). Urban shadow detection and classification using hyperspectral image. Journal of the Indian Society of Remote Sensing, 45(6), 945–952.CrossRefGoogle Scholar
  14. Renard, N., & Bourennane, S. (2008). Improvement of target detection methods by multiway filtering. IEEE Transactions on Geoscience and Remote Sensing, 46, 2407–2417.CrossRefGoogle Scholar
  15. Renard, N., Bourennane, S., & Blanc, J. (2008). Denoising and dimensionality reduction using multilinear tools for hyperspectral image. IEEE Geoscience and Remote Sensing Letters, 5, 138–142.CrossRefGoogle Scholar
  16. Tao, D., & Huang, X. (2013). Tensor discriminative locality alignment for hyperspectral image spectral-spatial feature extraction. IEEE Transactions on Geoscience and Remote Sensing, 51, 242–256.CrossRefGoogle Scholar
  17. Tong, Q., Zhang, B., & Zheng, L. (2006). Hyperspectral remote sensing: principles, techniques and applications. Beijing, China: Higher Education Press.Google Scholar
  18. Tucker, L. R. (1966). Some mathematical notes on three-mode factor analysis. Psychometrika, 31, 279–311.CrossRefGoogle Scholar
  19. Zhong, Z., Fan, B., Duan, J., Wang, L., Ding, K., Xiang, S., et al. (2015). Discriminant tensor spectral-spatial feature extraction for hyperspectral image classification. IEEE Geoscience and Remote Sensing Letters, 12, 1028–1032.CrossRefGoogle Scholar

Copyright information

© Indian Society of Remote Sensing 2018

Authors and Affiliations

  • Ronghua Yan
    • 1
    • 2
    Email author
  • Jinye Peng
    • 1
    • 3
  • Dongmei Ma
    • 4
  • Desheng Wen
    • 2
  1. 1.School of Electronics and Information, Space Optics LabNorthwestern Polytechnical UniversityXi’anPeople’s Republic of China
  2. 2.Xi’an Institute of Optics and Precision MechanicsCASXi’anPeople’s Republic of China
  3. 3.School of Information and TechnologyNorthwest UniversityXi’anPeople’s Republic of China
  4. 4.Xi’an-Janssen Pharmaceutical LtdXi’anPeople’s Republic of China

Personalised recommendations