一种改进的自适应加权中值滤波算法研究

Research on an Improved Adaptive Weighted Median Filtering Algorithm

图像在传输和处理过程中不可避免地存在边缘、细节信息破坏、垃圾噪声加载等问题,需要对输出图像进行过滤。对此,分析了几种典型的改进后的中值滤波算法,在此基础上提出了一种新的自适应加权中值滤波算法(WAMF)。该算法汲取了常见中值滤波算法的优缺点,通过噪声检测确定图像中的噪声点,根据窗口中噪声点的数量自适应调整滤波窗口的大小,像素点在滤波窗口中按照特定规律自适应分组,按照相似度值给各组像素点分配权重值,对检测到的噪声进行合理滤波。仿真结果表明,WAMF算法不仅可以有效去除噪声,而且较好地保存了图像细节,滤波性能优于常见中值滤波算法。

In the process of transmission and processing,there inevitably exist problems such as edge,detail information destruction,garbage noise loading and so on,so the output image needs to be filtered.In view of this,several typical improved median filtering algorithms are analyzed.On the basis,we propose a new adaptive weighted median filtering algorithm(WAMF).This algorithm captures the advantages and disadvantages of the common median filtering algorithm.The noise point in the image is determined by noise detection.The size of the filter window is adaptively adjusted according to the number of noise points in the window.Pixels are adaptively grouped in the filter window according to specific rules.According to the similarity value,weight values are assigned to each group of pixels,and the detected noise is properly filtered.The simulation shows that the WAMF algorithm not only can effectively remove the noise,but also preserves the details of the image better.Its filtering performance is better than the common median filtering algorithm.