脉冲噪声的存在,从视觉上影响图像的质量.为了去除图像中的脉冲噪声,提出了利用相似邻居数对图像中的噪声像素进行标记并建立相应矩阵,接着根据噪声的污染程度自适应的选择滤波窗口,最后对噪声像素进行自适应均值滤波去除图像中的脉冲噪声.实验结果表明,算法能有效降低图像中的脉冲噪声,对图像的边缘与细节保持较好.%The presence of impulse noise impacts from the visual quality of the image. A new denoising method is proposed based on the noise detection for removing and restoring impulse noise. The algorithm marks the noise of image pixel according to the similar number of neighbors,sets a marked matrix,adap-tively selects filter window according to the noise pollution degree and removes impulse noise in images u-sing the adaptive mean filter. Experimental results show that this algorithm can eliminate impulse noise and gives significantly better image quality than other algorithms such as mean filter, median filter and gradient inverse weighted filter. With the analysis of the experimental data,this method reaches a good result not only the details and edge information of the original image but also the PSNR.
展开▼