Abstract: Aiming at the problem of noise pollution in color image, this paper presents an improved adaptive switched vector filtering based on Peer Group Filtering(PGF). Using PGF detection noisy image to get the number of noise pixels which meet the test conditions in the filtering window, if the number of noise pixels which meet the test conditions is few, using directly vector median filtering for noise, if the number is large, using improved Adaptive Vector Median Filter(AVMF) to go on two time detection and decide filtering or not. Experimental results prove that the method has better performance on noise detection and protecting filtering details better.

Key words: color imag, impulse noise, Peer Group Filtering(PGF), Adaptive Vector Median Filter(AVMF)

摘要： 针对彩色图像中的噪声污染问题，提出一种改进的开关自适应矢量滤波方法。通过对噪声图像进行同组滤波器检测得到滤波窗口内满足检测条件的噪声像素个数，当满足条件的像素个数较少时，直接对检测出的噪声进行矢量中值滤波，当满足条件的像素个数较多时，采用改进的自适应矢量中值滤波器进行2次检测后再滤波。实验结果表明，该方法能提高噪声检测的准确性，并能更好保护滤波的细节。

关键词: 彩色图像, 脉冲噪声, 同组滤波器, 自适应矢量中值滤波器

CLC Number: 

• TP391.4