Speckle Filtering Algorithm of PolSAR Imagery Based on Two-Dimensional Polarimetric Subspace ICA

  • Hao-gui Cui
  • Gao-ming Huang
  • Tao Liu
  • Jun Gao
Conference paper


To improve the performance of speckle filter on polarimetric synthetic aperture radar (PolSAR) imagery, an ICA algorithm based on two-dimensional polarization subspace is proposed. In this method, the PolSAR data of three channels were divided into three two-dimensional subspaces, and then the speckle noise component and texture component can be separated by the ICA algorithm. The equivalent number of looks (ENL) can be used to evaluate the effect of speckle filter. And an automatic ENL estimation algorithm is introduced to avoid the manual selection of a region with fully developed speckle and no texture. Performance of the novel speckle filter is tested through real data and the results show that the proposed filter is effective and robust.


Independent component analysis Polarimetric subspace Polarimetric SAR Speckle 



This work was supported by the National Natural Science Foundation of China under Grant 61372165. The authors are also grateful to European Space Agency for providing the PolSAR data.


  1. 1.
    Novak LM, Burl MC (1990) Optimal speckle reduction in polarimetric SAR imagery. IEEE Trans Aerosp Electron Syst 26(2):293–305CrossRefADSGoogle Scholar
  2. 2.
    Lee JS, Grunes MR, Grandi G (1999) Polarimetric SAR speckle filtering and its implication for classification. IEEE Trans Geosci Remote Sens 37(5):2363–2373CrossRefADSGoogle Scholar
  3. 3.
    Chen HY, Li LM (2006) Speckle reduction in polarimetric SAR images based on independent component analysis. J Univ Electron Sci Technol China 35(4):480–483 (in Chinese)Google Scholar
  4. 4.
    Zhang ZS, Yu J, Yan Q, Zhao Z (2011) Research on polarimetric SAR image speckle reduction using kernel independent component analysis. Acta Geodaet Cartogr Sin 40(3):289–295 (in Chinese)MATHGoogle Scholar
  5. 5.
    Chen HY, Fu YS, Pi YM (2007) “Speckle reduction of polarimetric SAR image based on ICA-SCS algorithm. J Electron Inf Technol 29(4):819–821 (in Chinese)Google Scholar
  6. 6.
    Chen PS, Tong QX (1998) Remote sensing information mechanism research. Science press, Beijing, pp 43–50 (in Chinese)Google Scholar
  7. 7.
    Goodman N (1963) Statistical analysis based on a certain multivariate complex Gaussian distribution. Ann Math Stat 34(1):152–177 (an introduction)MATHCrossRefGoogle Scholar
  8. 8.
    Arsenault H, April G (1976) Properties of speckle integrated with a finite aperture and logarithmically transformed. JOSA 66(11):1160–1163CrossRefADSGoogle Scholar
  9. 9.
    Anfinsen SN, Doulgeris AP, Eltoft T (2009) Estimation of the equivalent number of looks in polarimetric synthetic aperture radar imagery. IEEE Trans Geosci Remote Sens 47(11):3795–3809CrossRefADSGoogle Scholar
  10. 10.
    Silverman BW (1986) Density estimation for statistics and data analysis. Chapman & Hall/CRC, Boca Raton, pp 103–108Google Scholar
  11. 11.
    Wand MP, Jones MC (1995) Kernel smoothing, Chapman & Hall/CRC, Boca Raton, pp 239–250Google Scholar
  12. 12.
    Hyvarinen A, Oja E (1997) A fast fixed-point algorithm for independent component analysis. Neural Comput 9(7):1483–1492CrossRefGoogle Scholar
  13. 13.
    Hyvarinen A (1999) Fast and robust fixed-point algorithms for independent component analysis. IEEE Trans Neural Netw 10(3):626–634CrossRefPubMedGoogle Scholar
  14. 14.
    Wang H, Pi Y, Liu G, Chen H (2008) Applications of ICA for the enhancement and classification of polarimetric SAR images. Int J Remote Sens 29(6):1649–1663CrossRefGoogle Scholar
  15. 15.
    Chitroub S, Amrane H, Boualem S (2000) Compound PCA-ICA neural network model for enhancement and feature extraction of multi-frequency polarimetric SAR imagery. In: Proceedings the 2000 International Conference on Image Processing, Vancouver, British Columbia, Canada, pp 328–331Google Scholar

Copyright information

© Atlantis Press and the author(s) 2016

Authors and Affiliations

  1. 1.School of Electronic EngineeringNaval University of EngineeringWuhanChina

Personalised recommendations