摘要
当图像中同时存在脉冲噪声和高斯噪声时,传统的中值滤波算法和均值滤波算法均不能达到较好的去噪效果。针对这一问题,提出了一种改进的加权均值滤波算法。算法采用局部阈值优化的方法计算各像素点的权值,将滤波窗口各像素点的灰度值与对应的权值进行加权运算,结果作为窗口中心点的滤波输出。仿真实验结果证明,该算法对脉冲噪声和高斯噪声具有较强的去噪能力,且较好地保持了图像的细节,效果均优于传统中值、均值滤波算法和改进的中值滤波算法(IMF)。
Traditional median filtering algorithm and mean filtering algorithm can’t remove the noise effectively for the image polluted by impulse noise and Gauss noise simultaneously. Aiming at this problem,an improved weighted mean filtering algo-rithm is presented. The algorithm takes local threshold optimization to calculate weight of each pixel in filtering windows,and then the weighted operation of gray value of each pixel in window and its corresponding weight is performed. The result is taken as the filtering output of center pixel in the window. Simulation experiment results demonstrate the new algorithm has strong de-noising ability for images polluted by impulse noise and Gauss noise,and preserves more image details. Its effect is better than traditional median filtering algorithms,mean filtering algorithms and even the improved median filtering algorithm(IMF).
出处
《现代电子技术》
北大核心
2015年第10期1-3,6,共4页
Modern Electronics Technique
基金
国家自然科学基金:基于博弈论的高效稳定聚类算法研究(61473045)
辽宁省高等学校实验室项目(L2012397)
博士后基金项目(2012M520158)
辽宁省"百千万人才工程"资助项目(2012921058)
教育厅科研一般项目(L2014451)
关键词
脉冲噪声
高斯噪声
均值滤波
中值滤波
impulse noise
Gauss noise
mean filtering
median filtering