一种基于中值分界的快速中值滤波算法

运用统计学原理和定位数值的快速查找算法提出一种基于中值分界的加速滤波算法．该算法以前次窗中数据的中值为支点将新窗口中数据分为大于、小于和等于该中值的三个集合，并统计各集合中数据个数，判断出中值所在集合的排序位，对集合该顺序住上的数据进行快速查找输出。实验证明，该算法大大降低了传统中值滤波的时间复杂度。提高运行速度，对于大量数据和大幅图像的滤波处理能更好地满足实时性要求。

A Kind of Fast Median Filtering Algorithm Based on Medium Value Separation

According to statistical principles and positioning numerical rapidly search algorithm, proposes a new kind of fast algorithm based on median demarcation. This method first takes the median data as a basis value. Then, according to the basis value, divide the new data in the window into three parts. The three sets are respectively less than the basis value, equal to the basis value and more than the basis value. After that, counts the number of elements in each set and judges the ranking position of that. After the previous steps, searches as well as compares the elements and outputs the data rapidly. Experimental result shows that, comparing with traditional median filtering method, this algorithm can greatly reduce the time complexity, improve the speed of data processing. This algorithm can also better satisfy the requirement in real-time performance for large amounts of data or large images.