Error Reduction in Images with the Use of Additional Information
Abstract
The use of additional information to achieve error reduction in images is discussed in this report. Theoretical results are derived for images corrupted with Fourier domain wide-sense stationary noise. Three types of additional information are explored: support, positivity, and high-quality prior image data. The basic mechanism for reducing noise in images with these types of additional information is shown to be enforced correlations in the Fourier data. The amount of noise reduction achieved is shown to be a function of the relative correlation of the noises and the enforced correlations. For significant noise reduction to be achieved, the noises must be significantly uncorrelated relative to the enforced correlations and the sizes of the noise variances must also be significantly uncorrelated. These results are applied to computer-simulated and field data from telescopes with adaptive optics.
Document Details
- Document Type
- Technical Report
- Publication Date
- Jan 31, 1994
- Accession Number
- ADA278636
Entities
People
- Charles L. Matson
Organizations
- Seattle Pacific University