一种新的加权中值滤波的快速算法

A New Fast Weighted Median Filtering Algorithm

针对中值滤波排序工作量多、速度慢等缺点，提出了一种新的加权中值滤波快速 算法。把窗口内元素分为小于、等于和大于中心元素的3类并分别统计这3类的个数，同时 构 造两个数组分别放入小于和大于中心元素的元素值，对每个类元素进行比较后可确定输 出中 值的确切位置，因此可根据局部排序找出输出中值，这样可大大减少排序的次数，加 快了处 理速度。最后，对快算算法和标准中值滤波的传统算法进行了比较，给出了二组实 验数据。 实验结果表明，与传统算法相比，该算法的执行速度大大加快。通过选择适当的 权值，加权 中值滤波可代表标准中值滤波和其它一些中值滤波，因此该快速算法有较强适 应性。

In order to solve the problem that the speed of median filtering is slow because of a lot of sorting, a new fast weighted median filtering algorit h m is presented. All elements in the moving window are classified as three kin ds by comparing with center element,and compute the number of each class respec tiv ely, and construct two new arrays that store those elements value that is sm alle r or bigger than center element. By several simple comparison, people can k now t he position of new median in new array. SO the number of sorting is decrea sed la rgely and the speed of filtering is improved.At last,two groups of test d ata are given by the comparison between the fast algorithm and the traditional algorith m. The result proves the fast algorithms to be efficient. Selecting app ropriate weight, weighted median filtering is equal to standard median filterin gs or some other median filterings,so the fast algorithms adaptability is large .