具有去除椒盐噪声能力的改进双边滤波算法

An improved bilateral filtering algorithm having the ability to remove salt-and-pepper noise

针对双边滤波器不能有效去除椒盐噪声这一特点,本文提出了一种自适应大小的模板替换算法,通过先替换双边滤波模板区域像素值,再对替换后的模板区域做双边滤波,使其在具有双边滤波保边并去除低频噪声能力的基础上兼具去除椒盐噪声的能力.实验结果证明:本文提出的算法在双边滤波的基础上可以有效去除图像中的椒盐噪声,与几种经典去除椒盐噪声的算法相比,在噪声密度10%~90%的情况下,峰值信噪比(PSNR)平均值提升5.8%以上,结构相似性(SSIM)平均值提升2.75%以上.

In allusion to the feature of bilateral filter having not the ability to remove salt-and-pepper noise,this paper proposes a adaptive template replacement algorithm,which firstly replaces the pixel value of bilateral filtering template areas and then bilateral filtering can be made for the replaced template areas to have the ability to remove the salt-and-pepper noise when equipped with the capacity of removing low-frequency noise while remaining edge with bilateral filtering.The experimental results show that improved bilateral filtering algorithm can remove salt-andpepper noise well,and compared with several classical algorithms having the ability to remove salt-and-pepper noise,the average value of PSNR calculated by the proposed algorithm has been improved more than 5.8%and the average value of SSIM has been enhanced more than 2.75% under the condition of 10%~90% of noise density.