Edge Detection Free Postprocessing for Pseudospectral Approximations

Article

Abstract

Pseudospectral Methods based on global polynomial approximation yield exponential accuracy when the underlying function is analytic. The presence of discontinuities destroys the extreme accuracy of the methods and the well-known Gibbs phenomenon appears. Several types of postprocessing methods have been developed to lessen the effects of the Gibbs phenomenon or even to restore spectral accuracy. The most powerful of the methods require that the locations of the discontinuities be precisely known. In this work we discuss postprocessing algorithms that are applicable when it is impractical, or difficult, or undesirable to pinpoint all discontinuity locations.

Keywords

Fourier Chebyshev Pseudospectral Gibbs Spectral filtering Digital total variation filtering 

References

  1. 1.
    Boyd, J.P.: Chebyshev and Fourier Spectral Methods, 2nd edn. Dover, New York (2000) Google Scholar
  2. 2.
    Canuto, C., Hussaini, M., Quarteroni, A., Zang, T.: Spectral Methods: Fundamentals in Single Domains. Springer, Berlin (2006) MATHGoogle Scholar
  3. 3.
    Chan, T., Osher, S., Shen, J.: The digital TV filter and nonlinear denoising. IEEE Trans. Image Process. 10(2), 231–241 (2001) MATHCrossRefGoogle Scholar
  4. 4.
    Gelb, A.: A hybrid approach to spectral reconstruction of piecewise smooth functions. J. Sci. Comput. 15, 293–322 (2001) CrossRefMathSciNetGoogle Scholar
  5. 5.
    Gelb, A., Cates, D.: Detection of edges in spectral data III: Refinement of the concentration method. J. Sci. Comput. 36(1), 1–43 (2008) CrossRefMathSciNetGoogle Scholar
  6. 6.
    Gelb, A., Tadmor, E.: Detection of edges in spectral data. Appl. Comput. Harmon. Anal. 7, 101–135 (1999) MATHCrossRefMathSciNetGoogle Scholar
  7. 7.
    Gelb, A., Tadmor, E.: Detection of edges in spectral data II: Nonlinear enhancement. SIAM J.  Numer. Anal. 38(4), 1389–1408 (2000) MATHCrossRefMathSciNetGoogle Scholar
  8. 8.
    Hesthaven, J., Gottlieb, S., Gottlieb, D.: Spectral Methods for Time-Dependent Problems. Cambridge University Press, Cambridge (2007) MATHGoogle Scholar
  9. 9.
    Hesthaven, J.S., Kirby, R.M.: Filtering in Legendre spectral methods. Math. Comput. 77, 1425–1452 (2008) CrossRefMathSciNetGoogle Scholar
  10. 10.
    Jung, J.-H., Shizgal, B.: Inverse polynomial reconstruction of two dimensional Fourier images. J. Sci. Comput. 25, 367–399 (2005) MATHCrossRefMathSciNetGoogle Scholar
  11. 11.
    Kosloff, R., Tal-Ezer, H.: A modified Chebyshev pseudospectral method with an O(1/n) time step restriction. J. Comput. Phys. 104, 457–469 (1993) MATHCrossRefMathSciNetGoogle Scholar
  12. 12.
    Ma, H.: Chebyshev-Legendre super spectral viscosity method for nonlinear conservation laws. SIAM J. Numer. Anal. 35, 893–908 (1998) MATHCrossRefMathSciNetGoogle Scholar
  13. 13.
    Osher, S., Shen, J.: Digitized PDE method for data restoration. In: Anastassiou, G. (ed.) Analytic-Computational Methods in Applied Mathematics, Chap. 16, pp. 751–771. Chapman and Hall/CRC (2000) Google Scholar
  14. 14.
    Sarra, S.A.: Digital Total Variation filtering as postprocessing for Chebyshev pseudospectral methods for conservation laws. Numer. Algorithms 41, 17–33 (2006) MATHCrossRefMathSciNetGoogle Scholar
  15. 15.
    Sarra, S.A.: Digital Total Variation filtering as postprocessing for Radial Basis Function Approximation Methods. Comput. Math. Appl. 52, 1119–1130 (2006) MATHCrossRefMathSciNetGoogle Scholar
  16. 16.
    Sarra, S.A.: The Matlab postprocessing toolkit. Submitted to ACM Trans. Math. Softw. (2009) Google Scholar
  17. 17.
    Sod, G.: A survey of several finite difference methods for systems of nonlinear hyperbolic conservation laws. J. Comput. Phys. 27, 1–31 (1978) MATHCrossRefMathSciNetGoogle Scholar
  18. 18.
    Trefethen, L.N.: Spectral Methods in Matlab. SIAM, Philadelphia (2000) MATHGoogle Scholar
  19. 19.
    Vandeven, H.: Family of spectral filters for discontinuous problems. SIAM J. Sci. Comput. 6, 159–192 (1991) MATHCrossRefMathSciNetGoogle Scholar

Copyright information

© Springer Science+Business Media, LLC 2009

Authors and Affiliations

  1. 1.Department of MathematicsMarshall UniversityHuntingtonUSA

Personalised recommendations