一种基于极值检测与自适应滤波的图像去噪算法

标准中值滤波及其一些改进算法对于被低密度脉冲噪声污染图像的处理可取得令人满意的效果,图像被严重污染时,这些算法得到的结果均不理想。针对这一缺陷,基于文献1]提出的一种滤波算法进行改进。首先,利用极值法对图像进行检测,判断出噪声及非噪声点;其次,设置滤波模板的最小及最大尺寸,对噪声点进行窗口逐渐增大的滤波处理。计算机模拟实验结果及对SAR图像滤波结果表明:该滤波算法在噪声去除及边缘和图像细节保持上优于标准中值滤波及其一些改进算法。

An effective filtering algorithm based on extremum detection and adaptive filtering

The standard median filter and some of its improved algorithms will work well when they are used to deal with images contaminated by low density impulse noises;however,results of these algorithms are not satisfactory when images are polluted densely.In reference to this limitation,an improved algorithm is presented in this paper.Firstly,detect noises using the method of extremum detection.Secondly,the filtering process is used to those noises with the size of filtering window augmenting automatically.Results of the computer simulation experiments and the processed SAR images show that noise removal and edge or detail maintenance of an image of this paper’s filter method are better than the standard median filter and some of its improved algorithms.