自适应中心加权的彩色图像中值滤波方法

Median filtering method based on adaptive central weighting for color images

提出一种彩色图像自适应中心加权的矢量中值滤波方法.通过引入投影距离确定与待滤波点最接近的若干个像素的关系,并考虑该点与由聚合距离所确定的中值的关系,自适应地确定该点的权系数,形成相应模板下的一个多重向量集合,将该集合的中值作为滤波中值.实验证明所提出方法能有效地去除彩色图像的椒盐噪声,同时能够较好地保持其色调和细节信息.

An adaptive center weighted vector median filtering for color images is proposed.The relationship between the filtered pixel and its several nearest neighbor pixels is determined by introducing the projective distance.The link between the given pixel and the median value of template based on the aggregative distance is considered,and the weighting coefficient of the given pixel is adaptively obtained.A multi-set of vectors is formed and the median value of the multi-set is taken as the median of filtering.Experimental results show that the proposed method can effectively remove salt and pepper noises in a color images and preserve color information and local details.