Smoothing filter for color image based on Karhunen-Loeve transform and recursive least-square
Shi Chong-shen · Jisuanji gongcheng yu sheji · 2015
Aiming at the problems of the smoothing filter for color image,an image smoothing method based on Karhunen-Loeve(KL)transform and recursive least-square(RLS)was presented.The color image was transformed from the RGB color space into a new space(JKL space)using KL transform,and then filtering method of RLS was used to smooth component images.The smoothed component images were synthesized and transformed from the JKL space into RGB space using KL inversion transform.Compared with restoration effects of the constant ratio Wiener filtering and auto correlation function Wiener filtering,results show that the proposed method has good smoothing effect for the blurred color image with Gaussian noise and salt and pepper noise added.