一种基于改进粒子群的图像滤波方法

摘要

In light of the deficiencies of traditional median filter, in this paper we propose to adopt an improved PSO-random inertia weight method to seek the optimal parameter of fuzzy membership function, and then perform fuzzy median filtering on the noise image as well as recover the gray value of the pixel points stained by noise. Experimental results using Matlab image processing toolbox indicate that the effect of median filter based on improved PSO has certain improvement than the effect of general median filter. The method has quite good adaptability and can continuously approach the target, it can also well protect the detailed information of the images and improve the denoising effect and the clarity of the images.%根据传统中值滤波的不足,提出采用改进的粒子群算法PSO(Particle Swarm Optimization)-随机惯性权法来寻求模糊隶属函数的最优参数,然后对噪声图像进行模糊中值滤波处理,复原被噪声污染的像素点灰度值.通过Matlab图像处理工具箱的实验结果表明,基于改进PSO中值滤波比一般中值滤波效果有了一定程度的提高.该方法不仅有较好自适应性,可以不断逼近目标值,而且还能很好地保护图像细节信息,提高图像的去噪效果和清晰度.