摘要
提出的改进的极值中值滤波(IEM)算法是在极值中值滤波(EM)算法的基础上做了如下两点改进。首先,采用更合理的检测方法来检测噪声点,减小将信号点误判为噪声点的概率。其次,采用改进的滤波算法,解决了当噪声点个数大于像素总数一半时,传统中值滤波无能为力的问题。实验证明,当噪声密度很大时,IEM方法不仅比EM方法有更好的滤波效果,而且能够更好地保护图像细节。
The improved extremum and median （IEM） filter was proposed which is based on the extremum median （EM） filter. It is improved in two sides. For one thing, in order to reduce the possibility of mistaking the signal pixels for noise pixels, the more appropriate method is used to detect noise pixels. For another, when noise pixels are more than half of the total pixels, the standard median filtering algorithm can not resolve it. IEM adapts the improved filtering algorithm to solve it. Experimental results show that when the density of noise is high, the IEM filtering not only has better performance than EM filtering, but also keeps more details than EM filtering.
出处
《计算机科学》
CSCD
北大核心
2009年第6期165-166,共2页
Computer Science
关键词
改进的极值中值滤波
噪声检测
阈值
Improved extremum and median filtering, Noise detection, Threshold
