改进的快速中值滤波算法在图像去噪中的应用

摘要

In the image processing, the traditional median filtering algorithm is based on sorting algotithm, the correlation of datas are not considered. In the paper, we will list an improved fast algorithm of median filter; the correlation of datas are full used in this algorithm; we don’t need comparing many datas. First window's median value can be computed by divide-and-conquer method, then use the correlation of datas to compute the left windows's median values. This method can greatly improve the speed of median filtering. At last, we will use traditional median filtering algorithm and fast algorithm of median filter to deal with the same image with salt-and-pepper noise. It can conclude that the fast algortihtm of median filter can remove image noise quickly by the trial.%在图像处理中,传统的中值滤波算法都是建立在排序理论上,并没有充分考虑到各数据之间的相关性.提出一种改进的快速中值滤波算法,充分利用数据的相关性,避免传统算法因排序所需大量数据比较.先用分治法计算第一个窗口的中值,然后利用数据的相关性计算余下窗口的中值,大大提高了效率.最后分别用传统的中值滤波算法和改进的快速中值滤波算法对同一幅有椒盐噪声的图像做去噪实验.实验表明,该算法能快速实现图像去噪.