Parameter Estimation for the Blind Restoration of Blurred Imagery.
Abstract
Methods to estimate the optical transfer function (OTF) of the degradation and the noise-to-signal ratio in a degraded image are presented which use no prior information about the degradation and the noise. Two types of degradations are considered: uniform linear camera motion and out-of-focus camera with circular aperture. The OTF of the out-of-focus camera is estimated by inspection of the spectra of the degraded image and from average over the spectra of subsections of the degraded image. A line estimation method is described to estimate the degrading OTF of the linear camera motion in one direction. An additional method called the logarithmic estimate is also developed. Frequency and space domain estimation methods are developed for the noise-to-signal ratio for the case of additive gaussian noise. Inverse, Wiener, and power spectrum equalization filters based on these estimates are implemented on a digital computer to restore a variety of degraded images and examples of blind restoration of these degraded images are presented. (Thesis).
Document Details
- Document Type
- Technical Report
- Publication Date
- Sep 01, 1986
- Accession Number
- ADA177602
Entities
People
- Murat Cincioglu
Organizations
- Air Force Institute of Technology