一种基于噪声点检测的自适应中值滤波方法  

An adaptive median filtering algorithm based on noise detection

作　　者：高克芳[1] 郭建钢[1] 

机构地区：[1]福建农林大学机电工程学院,福建福州350002

出　　处：《福建农林大学学报（自然科学版）》2009年第3期333-336,共4页Journal of Fujian Agriculture and Forestry University:Natural Science Edition

基　　金：国家自然科学基金资助项目(30370393)

摘　　要：针对中值滤波在图像去噪时会造成图像细节丢失的问题,提出了一种新的基于噪声点检测的自适应中值滤波法.该方法对噪声点采用两级判断的方法:首先根据椒盐噪声的特点将图像像素点分为可疑噪声和信号两类;对于可疑噪声点,根据噪声与细节在图像中的表现,将可疑噪声分为噪声和边缘细节;然后采用不同的中值滤波窗口对噪声点进行滤波,对于两次判断得到的信号和边缘细节不进行处理以保持图像的细节.测试结果表明,与常用的中值滤波法相比,该方法不仅具有较好的去噪特性,还具有较强的细节保护能力.To overcome the drawback that median filtering algorithm often get rid of fringe details of image while filtering noise, the adaptive median filtering algorithm based on noise detection is proposed in this paper. The algorithm applys the two stages method to detect noise. First, the algorithm determines a pixel point as signal pixel or possible noise pixel according to the characteristic of salt and pepper noise. Second, according to the difference between salt and pepper noise and fringe details, each possible noise pixel is classified to be noise pixel or fringe details. Finally, the algorithm employs adaptive median filtering to remove the noise pixels, while the signal pixels and fringe detail pixels are kept untouched in the filtering process to preserve the detail of image. Simulation results demonstrate that this algorithm is better than traditional median-based fihers in performance of filtering noise and detail preservation.

关 键 词：图像处理 中值滤波 自适应 噪声检测 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]