一种有效的自适应中值滤波算法

An effective adaptive median filter algorithm

针对标准中值滤波算法的不足,提出改进的自适应中值滤波算法。该算法利用极值的方法统计图像噪声点,然后计算以噪声点为中心像素的子窗口的方差,对噪声点采取多窗口、多尺度的滤波。最后自适应选择方差最小的子窗口滤波。经过计算机模拟实验,采用该算法滤波,既能有效地平滑噪声,又能保存细节,效果远优于标准中值滤波算法。

Aim at the shortcomings of the standard median filter, we introduce an improved adaptive median filter algorithm which is based on the extremums to count the total numbers of noise and then calculate the variance of sub-windows around the noise point. Multi-scale and multi-window filter is applied to the noise points and at last the sub-window with the minimal variance is chosen. The results of computer simulation show that the algorithm can effectively suppress the noises but keep the details of the image. It is better than the standard median filter.