A generalized Gauss Markov model for space objects in blind restoration of adaptive optics telescope images

Author

Jeffs, Brian D. ; Hong, Sheila ; Christou, Julian

Author_Institution

Dept. of Electr. & Comput. Eng., Brigham Young Univ., Provo, UT, USA

fYear

1998

fDate

4-7 Oct 1998

Firstpage

737

Abstract

This paper introduces a blind method based on Bayesian maximum a posteriori estimation theory for restoring images corrupted by noise and blurred by one or more unknown point spread functions. Image and blur prior information is expressed in the form of parametric generalized Gauss Markov random field models. A method for estimating the GGMRF neighborhood influence parameters is presented, along with examples of blind restoration to reduce residual blur in adaptive optics telescope images of space objects

Keywords

Bayes methods; Gaussian processes; Markov processes; adaptive optics; astronomical techniques; image enhancement; image restoration; maximum likelihood estimation; optical transfer function; parameter estimation; Bayesian maximum a posteriori estimation theory; GGMRF neighborhood influence parameters; adaptive optics telescope images; blind restoration; blur; generalized Gauss Markov model; noise; parametric generalized Gauss Markov random field models; prior information; space objects; unknown point spread functions; Adaptive optics; Bayesian methods; Gaussian processes; Image restoration; Markov random fields; Maximum a posteriori estimation; Optical noise; Pixel; Probability density function; Shape; Telescopes;

fLanguage

English

Publisher

ieee

Conference_Titel

Image Processing, 1998. ICIP 98. Proceedings. 1998 International Conference on

Conference_Location

Chicago, IL

Print_ISBN

0-8186-8821-1

Type

conf

DOI

10.1109/ICIP.1998.727363

Filename

727363