Details

REGULARIZATION METHODS FOR ILL-POSED POISSON IMAGING

by Laobeul, N'Djekornom Dara

Abstract (Summary)

The noise contained in images collected by a charge coupled device (CCD) camera is predominantly of Poisson type. This motivates the use of the negative logarithm of the Poisson likelihood in place of the ubiquitous least squares t-to-data. However, if the underlying mathematical model is assumed to have the form z = Au, where A is a linear, compact operator, the problem of minimizing the negative log-Poisson likelihood function is ill-posed, and hence some form of regularization is required. In this work, it involves solving a variational problem of the form u def = arg min u0 `(Au; z) + J(u); where ` is the negative-log of a Poisson likelihood functional, and J is a regularization functional. The main result of this thesis is a theoretical analysis of this variational problem for four dierent regularization functionals. In addition, this work presents an ecient computational method for its solution, and the demonstration of the eectiveness of this approach in practice by applying the algorithm to simulated astronomical imaging data corrupted by the CCD camera noise model mentioned above.

Bibliographical Information:

Advisor:Jesse Johnson; Brian Steele; Leonid Kalachev; Thomas Tonev; Johnathan Bardsley

School:The University of Montana

School Location:USA - Montana

Source Type:Master's Thesis

Keywords:mathematical sciences

ISBN:

Date of Publication:01/15/2009

© 2009 OpenThesis.org. All Rights Reserved.