Nouvelle méthode de restauration d’images photographiques en couleur dégradées par le bruit de la transformation RGB→YCbCr

This paper presents a new method of restoration of photographic color images corrupted by the noise of transformation RGB→YCbCr. This transformation introduces into certain images a high frequency type noise. This noise is highlighted and corrected. The correction is made by attenuation of the coefficients Y, Cb and Cr followed by their rebuilding by multiplicative factors, which come from transformation relations analysis of spaces RGB and YCbCr. A comparative study with the Wiener’s filter shows its effectiveness to restore images corrupted even by other noises like the black caused by an insufficiency light during snapping process and its superiority on the median filter as for the correction of this noise.