Hostname: page-component-848d4c4894-p2v8j Total loading time: 0.001 Render date: 2024-05-13T20:54:28.023Z Has data issue: false hasContentIssue false
  • English
  • Français

The Effects of Errors on the Convergence of an Iterative Deconvolution Method*

Published online by Cambridge University Press:  06 March 2019

H. H. Madden  and
J. E. Houston
H. H. Madden
Affiliation:
Sandia Laboratories Albuquerque, New Mexico 87115
J. E. Houston
Affiliation:
Sandia Laboratories Albuquerque, New Mexico 87115
Get access

Abstract

Calculations carried out to investigate the van Cittert iterative deconvolution method and the effects of random noise and truncation errors on its convergence behavior are presented. Gaussian functions are used for "both the true function W and the system, response function A. The model "observed" function S is generated from W and A. Both rms differences between the result of n iterations Wn and the true function, and between Wn*A and S are used to measure convergence. The effects of introducing errors can be measured against standards set by the convergence without such errors. These "no-error" calculations make use of true functions with widths from 2.0 to 0.67 times the response function width. The effects of random noise are investigated by adding noise to W*A before deconvolution. A small amount of random noise initially added builds up rapidly in amplitude during the iterative process and eventually dominates the rms difference calculations. To suppress the effects of random noise build-up; smoothing techniques are applied, the best of which involved smoothing both the noisy observed function, S, and the system response function, A, before deconvolution. The smoothing operation is thus taken as part of the measurement and. the divergence resulting from, the noise build-up is avoided. The results depend strongly upon, the -width of the smoothing function. Uhsymmetric system response functions, similar' to those encountered in soft x-ray appearance potential spectroscopy and in x-ray continuum, isochromat measurements, are used in investigations of truncation errors. Abrupt cut-offs of the model S and A functions before deconvolution result in the build-up of large fluctuations in W . These truncation errors become increasingly localized with continued iterations and make only minor contributions to the errors in Wn in the vicinity of the real peak if the truncations are made sufficiently far from the peak location. Alternatively, the truncation errors can be avoided by analytical continuation.

Type
Soft X-Ray and Surface Analysis
Information
Advances in X-Ray Analysis , Volume 19: Twenty-Fourth Annual Conference on Applications of X-ray Analysis, August 6-8, 1975 , 1975 , pp. 657 - 671
Copyright
Copyright © International Centre for Diffraction Data 1975

Access options

Get access to the full version of this content by using one of the access options below. (Log in options will check for institutional or personal access. Content may require purchase if you do not have access.)

Footnotes

*

This work supported by the U.S. Energy Research and Development Administration.

References

1. Van Cittert, P. H., “ZumEinfluss der Spaltbreite auf die Intensitätsverteilung in Spektrallinien.II.” Z. Physik 69 293–308 (1931).Google Scholar
2. Burger, H. C. and van Cittert, P. H., “Wahre und Scheinbare Intensitätsverteilung in Spektrallinien,” Z. Pnysik 79, 722–730 (1932).Google Scholar
3. Bracewell, R. N. and Roberts, J. A., “Aerial Smoothing in Radio Astronomy,” Australian J. Phys. 7, 615–640 (1954).Google Scholar
4. Stokes, A. B., “A Numerical Fourier-Analysis Method for the Correction of Widths and Shapes of Lines on X-ray Powder Photographs,” Proc. Phys. Soc. (London) 61A, 382–391 (1948).Google Scholar
5. Kreisel, G., “Some Remarks on Integral Equations with Kernels: L (ζl-x1,…,ζn-xn; α),” Proc. Roy. Soc. (London) 197A, 160–183 (1949).Google Scholar
6. Fellgett, P. B. and Schmeidler, F. B., “On the Sharpening of Observational Data with Special Application to the Darkening of the Solar Limb,” Roy. Astron. Soc., Monthly Notices 112, 445–451(1952).Google Scholar
7. Burr, E. J., “Sharpening of Observational Data in Two Dimensions,” Australian J. Phys. 8, 30–53 (1955).Google Scholar
8. Braeewell, R. N., “Restoration in the Presence of Errors,” Proc. IRE 46, 106–111(1958).Google Scholar
9. Porteus, J. O., “Optimized Method for Correcting Smearing Aberrations: Complex X-ray Spectra,” J. Appl. Phys. 33, 700–707(1962).Google Scholar
10. Morrison, J. D., “On the Optimum Use of Ionization-Efficiency Data,” J. Chem. Phys. 39, 200–207 (1963).Google Scholar
11. Sauder, W. C., “General Method of Treating Instrumental Distortion of Spectral Data with Applications to X-ray Physics,” J. Appl. Phys. 37, 1495–1507 (1966).Google Scholar
12. Jones, A. F. and Misell, D. L., “The Problem of Error in Deconvolution,” J. Phys. A: Gen.Phys, 3, 462–472 (1970).Google Scholar
13. Misell, D. L. and Childs, P. A., “Deconvolution in Two Dimensions with Particular References to the Electron Microscope,” J. Phys. D: Appl. Phys. 5, 1760–1768 (1972).Google Scholar
14. Zenitani, F. and Minami, S., “An Analysis of the Iterative Method for Deconvolving Spectroscopic Data Containing a Random Noise,” Jap. J. Appl. Phys. 12, 379–387 (1973).Google Scholar
15. den Harder, A. and de Galan, L., “Evaluation of a Method for RealTime Deconvolution,” Anal. Chem. 46, 1464–1470 (1974).Google Scholar
16. Wertheim, G. K., “Deconvolution and Smoothing: Application in ESCA,” J. Electron Spectr. 6, 239–251 (1975).Google Scholar
17. Complementary results to those reported here will be presented in a separate report to be published.Google Scholar
18. Savitzky, A. and Golay, M. J. E., “Smoothing and Differentiation of Data by Simplified Least Squares Procedures,” Anal.Chem, 36 1627–1639 (1964).Google Scholar