摘要
针对传统自适应中值滤波算法的不足,文中提出了一种改进的自适应中值滤波方法,以有效的去除图像中的高密度脉冲噪声。第一,对于噪声点的检测,首先利用极大值和极小值的数量差找出可疑的噪声点,再利用邻域像素的相似性判断可疑点是否为噪声点。第二,对于滤波中值的计算,先把滤波窗口内具有相同灰度值的极值点压缩到一个,然后再计算中值。实验结果表明,该算法的滤波效果优于传统自适应中值滤波,且具有较好的稳定性。
To overcome the drawbacks of the traditional adaptive median filter, an improved adaptive medi- an filter is proposed to effectively remove the high density impulse noise of the image in this paper. Firstly, detect noise points. Different point values between the maximum and minimum points are used to identify suspicious noises. Then, the similarity of neighborhood pixels is used to judge these suspicious points whether they are noise points or not. Secondly, calculate the filtering median. Extreme points that have the same gray value in the filter window are compressed to only one. Finally, calculate the median values. The experimental result shows that this algorithm is better than traditional adaptive median filter to remove the noise and has a good stability.
出处
《通信技术》
2014年第8期873-876,共4页
Communications Technology
基金
国家自然科学基金(No.11204046)
贵州省科技厅工业攻关项目(黔科合GY字[2010]3056)
贵州大学研究生创新基金(研理工2014007)资助
关键词
中值滤波
脉冲噪声
自适应
极值
median filter
impulse noise
adaptive
extreme value