应用极值点检测的自适应中值滤波

Adaptive Median Filtering Based on Extreme Pixel Detection

针对图像中存在的椒盐噪声,提出一种新的基于极值点检测的自适应中值滤波方法。该方法通过对极值点的检测确定图像中可能存在的噪声点,并标记噪声点;然后采用窗口尺度可伸缩的自适应中值滤波方法对噪声点进行滤波。实验结果表明:自适应中值滤波方法在有效去除椒盐噪声的同时很好地保护了图像细节。通过与一些其他改进滤波算法的比较,证明了该方法具有更好的性能。

A new adaptive filter is proposed based on extreme pixel detection,in the purpose of solving the image pepper noise. In the algorithm,extreme pixels are detected as suspect noise pixels,and labeled as noise pixel map.Then scalable windows are selected adaptively in order to find proper pixel for filtering the noise pixels with median filter.Results of computer simulation experiments on the adaptive median filter show that the method can suppress pepper noise effectively while preserving fine details very well and has better filtering performance than the other advanced filters.