A Study of the Maximum Entropy Method of Power Spectrum Estimation as Applied to Interferometer Data.

Abstract

Maximum entropy method computer experiments were conducted on synthetic CO2 and OH input signals. They compared the power spectra which MEM produced with the given, exactly known, line spectra as well as with spectra produced by transitional Fourier methods. ARC scientists concluded from their study that MEM produces power spectra which are characterized by: (1) far higher resolution (narrower lines widths) than that produced by Fourier methods, (2) errors in frequency (line positions) which are about 2 orders of magnitude smaller than those of Fourier methods, and (3) when the frequencies have been determined by locating maxima in the MEM aor Fourier spectra, finding the amplitudes is a linear least squares problem. The errors in the MEM amplitude are about two orders of magnitude smaller than those of the Fourier methods. In the case of the quasiregular spectra, the implied very high resolution and accuracy provided by the MEM is contingent on an interferogram length equal to at least the reciprocal of the minimum line separation, and equivalent to at least four independent samples for each spectrum line. Keywords: Hydroxyl, Radicals, Carbon Dioxide; Maximum Entropy Method; Power Spectra; Interferometer Data.

Document Details

Document Type
Technical Report
Publication Date
Nov 17, 1986
Accession Number
ADA179403

Entities

People

  • A. S. Zachor
  • Douglas Smith

Organizations

  • Utah State University

Tags

Communities of Interest

  • Energy and Power Technologies

DTIC Thesaurus Topics

  • Amplitude
  • Carbon Dioxide
  • Errors
  • Frequency
  • High Resolution
  • Interferograms
  • Interferometers
  • Line Spectra
  • Power Spectra
  • Spectra

Fields of Study

  • Physics

Readers

  • Approximation Theory.
  • Image Processing and Computer Vision.
  • Molecular Photonics/Laser Physics