基于自适应中值滤波算法的研究与改进  

Research and Improvement of Adaptive Median Filtering Algorithm

作　　者：赵玲玲 ZHAO Ling-ling(College of Computer and Information Science,Chongqing Normal University,Chongqing 401331)

机构地区：[1]重庆师范大学计算机与信息科学学院,重庆401331

出　　处：《现代计算机》2019年第8期58-62,共5页Modern Computer

摘　　要：中值滤波器受滤波窗口大小影响较大,与中值滤波相比较,自适应中值滤波可以在一定程度上很好地抑制图像的脉冲噪声。然而,随着噪声密度的增加,自适应中值滤波算法的滤波效果降低。对于这两种滤波算法中的问题,提出一种基于自适应中值滤波算法的改进算法。该算法通过对滤波窗口中的图像灰度极值进行噪声检测,用滤波窗口的灰度中值代替噪声点。若滤波窗口的中值为噪声点,则算法自适应地增大滤波窗口从而取新的中值。如果滤波窗口自适应地增大到允许的最大窗口尺寸时,中值依然为噪声点,则取当前窗口中除去极大和极小像素的灰度均值。仿真实验结果表明,在随着噪声密度增大的情况下,提出的自适应中值滤波的改进算法滤波性能优于中值滤波算法和标准的自适应中值滤波。The median filter is greatly affected by the size of the filter window. Compared with the median filter, the adaptive median filter can suppress the impulse noise of the image to a certain extent. However, with the increase of noise density, the filtering effect of the adaptive median filtering algorithm gradually decreases. Aiming at the defects of the median filtering algorithm and the adaptive median filtering algorithm, proposes an improved algorithm of the adaptive median filtering algorithm. The algorithm detects the noise from the image grayscale extremum in the filter window, and replaces the noise point with the gray-scale median of the filter window. If the median of the filter window is the noise point, the algorithm adaptively increases the filter window to take a new median. If the median is still the noise point when the filter window adaptively increases to the maximum window size allowed, the gray-scale average of the maximum and minimum pixels removed from the current window is taken. Simulation results show that the improved adaptive median filtering algorithm has better filtering performance than the median filtering algorithm and the standard adaptive median filtering algorithm when the noise density increases.

关 键 词：中值滤波 噪声检测 自适应中值滤波 灰度极值 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术] TN713[自动化与计算机技术—计算机科学与技术]