Telecommunications and Radio Engineering

ISSN for PRINT: 0040-2508

For Online Access

Best Paper Award Selection - Editorial Board Site

2006, Volume65

Issue 6-10

470 pages

DOI: 10.1615/TelecomRadEng.v65.i6

Issue price - $1185.00

V. V. Lukin
Department of Transmitters, Receivers and Signal Processing, National Aerospace University, Chkalova Str. 17, 61070 Kharkiv, Ukraine

S. K. Abramov
Department of Transmitters, Receivers and Signal Processing, National Aerospace University, Chkalova Str. 17, 61070 Kharkiv, Ukraine

N. N. Ponomarenko
Department of Transmitters, Receivers and Signal Processing, National Aerospace University, Chkalova Str. 17, 61070 Kharkiv, Ukraine

Benoit Vozel
University of Rennes I, 6, Rue de Kerampont, 22 305 Lannion cedex, BP 80518, France

Kacem Chehdi
University of Rennes I, 6, Rue de Kerampont, 22 305 Lannion cedex, BP 80518, France

A priori knowledge of noise type and statistical characteristics (at least, noise variance) is a pre-requisite of successful solving many practical tasks of image processing like filtering (denoising), edge detection, segmentation, etc. This relates to both gray-scale and multichannel (multi- and hyper-spectral) images. However, in practical situations noise characteristics depend upon many factors and it is a common practice to determine noise type and variance just for images at hand. It is desirable to perform this operation automatically especially if one deals with multichannel images. Thus, below we consider several approaches and methods for blind evaluation of noise variance. Their accuracy and applicability are analyzed for images with different properties, in particular, various percentage of texture regions. Besides, a set of noise variance values is tested.

DOI: 10.1615/TelecomRadEng.v65.i6.40 Download article, 527-556 pages

Article price - $35.00

<< Previous article   Next article >>