一种基于改进的非对称裁剪中值滤波算法

A Median Filter Algorithm Based on Improved Unsymmetrical Trimmed

为了在各种噪声密度条件下,都能恢复椒盐噪声污染的图像并能很好地保持图像的细节,提出了一种基于改进的非对称裁剪中值滤波算法清除椒盐噪声。该方法首先对噪声点进行检测,然后基于滑动窗口中噪声点的数目来自适应改变窗口的大小,最后应用一种改进的非对称裁剪中值滤波器计算中值,结果显示该算法各项指标都要优于其它算法。实验结果表明了在各种的噪声密度条件下,该算法能较好地清除椒盐噪声,而且也能较好地保护图像细节,比现存的一些中值滤波算法清除椒盐噪声的效果更加优秀。

In order to restore the image as well as preserve image details with salt-pepper noise under a wide range of noise densities, pro- pose a median filter algorithm based on improved unsymmetrical trimmed for removal of salt-pepper. This way firstly check the salt-pep- per noise,then according to the number of noises in the window adaptively adjust the filter window size, finally get median by an im- proved unsymmetrical trimmed median filter. The results show that this algorithm is better than other algorithms in each index. The experi- mental results show that this algorithm may better eliminate salt-pepper noise and effectively protect image detail well, and it has a out- standing result for removal salt-pepper in comparison with other existing median algorithms under all kinds of noise densities.