高密度噪声和混合噪声的图像去噪算法

Images Denoising Algorithm for High Density Noise and Mixed Noise Removal

去除高密度噪声和混合噪声一直是图像处理的难题之一.提出的算法MMM(Mixed-Median-Mean)将传统的中值和均值滤波相结合,取得了较好的去噪效果.该算法对邻域窗口内像素点进行有效替换,并且引入超参数k对替换数目进行控制,有效地提升了传统算法的去噪性能.实验对单一的高密度噪声和混合噪声的不同组合进行了验证,噪声包括高斯噪声、脉冲噪声和乘性噪声.实验结果证明该算法的有效性及强鲁棒性.

Eliminating high density noise and mixed noise from the image is one of the difficult problems in image processing.The proposed algorithm combining the traditional Median and Mean filtering algorithm,called MMM(mixed-median-mean),achieves a better denoising effect.The algorithm replace the pixels in the processing window with appropriate values and control the replacement number by introducing a hyper parameter k.The experiments simulate single high-density noise,including gaussian noise,impulse noise and multiplicity noise,and the different combinations of mixed noise.The experimental results show that the MMM algorithm effectively improves the performance of the traditional algorithms and it is effective and robust.