【摘要】 基于对噪声模型的分析 ,提出了两种自适应中值滤波器算法 :(1)对于消除噪声密度较大的脉冲干扰 ,提出了基于剩余脉冲检测的自适应中值滤波器算法 (AMF1) ,通过对滤波器输出是否存在剩余噪声的检测来决定采用的滤波器窗口尺寸 ,从而有效滤除脉冲干扰 ;(2 )对于消除的一定宽度的脉冲干扰 ,则提出了基于脉冲宽度检测的自适应中值滤波器算法 (AMF2 ) ,通过决定干扰脉冲的宽度来确定滤波器窗口的大小 ,从而有效去除较宽的干扰脉冲。通过对实际图像的测试 ,表明提出的算法运算结果优于标准中值滤波器的输出结果更多还原

【Abstract】 Based on examining of the noise pattern, two kinds of adaptive median filter algorithms were proposed in this paper. (1)To remove impulse noise that have high occurrence probability, we have adaptive median filter algorithm based on detecting of residual impulses (note as AMF1), through detecting the presence of residual impulse in the median filter output to decide the window size of filter, therefore effectively remove impulse noise;(2) To remove impulse noise that have certain width, we have adaptive median filter algorithm based on detecting of the width of impulse (note as AMF2), through detecting the width of impulse noise to decide the size of filter, then we can remove impulse noise effectively. The algorithms had been tested on actual images, the results show that they are superior to the output of standard median filter.更多还原