Bayesian estimation of abrupt changes contaminated by multiplicative noise using MCMC

Author

Tourneret, Jean-Yves ; Doisy, Michel ; Mazzei, Manuel

Author_Institution

ENSEIHT/GAPSE, Nat. Polytech. Inst. of Toulouse, France

Volume

4

fYear

1998

fDate

12-15 May 1998

Firstpage

2133

Abstract

The paper addresses the estimation of abrupt changes which are contaminated by multiplicative Gaussian noise. The marginal mean a posteriori or marginal maximum a posteriori estimators can be derived for estimating the position of a single abrupt change. However, these estimators have optimization or integration problems for multiple abrupt changes. The paper solves these optimization problems by using Markov chain Monte Carlo methods (MCMC)

Keywords

Bayes methods; Gaussian noise; Markov processes; Monte Carlo methods; maximum likelihood estimation; parameter estimation; radar detection; radar imaging; spectral analysis; synthetic aperture radar; white noise; Bayesian estimation; MCMC; Markov chain Monte Carlo methods; SAR image processing; abrupt change estimation; integration problem; marginal maximum a posteriori estimator; marginal mean a posteriori estimator; multiplicative white Gaussian noise; optimization problem; position estimation; spectral analysis algorithm; Bayesian methods; Gaussian noise; Image processing; Maximum a posteriori estimation; Maximum likelihood estimation; Optimization methods; Reflectivity; Signal processing; Speckle; Synthetic aperture radar;

fLanguage

English

Publisher

ieee

Conference_Titel

Acoustics, Speech and Signal Processing, 1998. Proceedings of the 1998 IEEE International Conference on

Conference_Location

Seattle, WA

ISSN

1520-6149

Print_ISBN

0-7803-4428-6

Type

conf

DOI

10.1109/ICASSP.1998.681567

Filename

681567