Least-squares interband denoising of color and multispectral images

Author

Scheunders, P. ; Driesen, J.

Author_Institution

Dept. of Phys., Antwerp Univ., Belgium

Volume

2

fYear

2004

fDate

24-27 Oct. 2004

Firstpage

985

Abstract

This paper exploits the interband correlations of color and multispectral images for wavelet-based denoising. For this, a multispectral extension of the linear minimum mean squared error estimation (LMMSE) is constructed to estimate the signal from the observed wavelet coefficients. The calculation involves the signal autocovariance matrices, which are estimated globally or locally far centered square windows using maximum likelihood and MAP. The method is demonstrated to outperform single-hand denoising on color and 7-band Landsat multispectral images.

Keywords

correlation theory; covariance matrices; image colour analysis; image denoising; image reconstruction; least mean squares methods; maximum likelihood estimation; wavelet transforms; 7-band Landsat multispectral image; LMMSE; MAP; centered square window; global estimation; interband color correlation; linear minimum mean squared error estimation; maximum a posteriori estimation; maximum likelihood; signal autocovariance matrix; signal construction; single-hand denoising; wavelet-based denoising; Additive noise; Discrete wavelet transforms; Gaussian noise; Multispectral imaging; Noise reduction; Physics; Remote sensing; Signal to noise ratio; Spatial resolution; Wavelet transforms;

fLanguage

English

Publisher

ieee

Conference_Titel

Image Processing, 2004. ICIP '04. 2004 International Conference on

ISSN

1522-4880

Print_ISBN

0-7803-8554-3

Type

conf

DOI

10.1109/ICIP.2004.1419466

Filename

1419466