带边缘增强的反比相对距离加权均值滤波

Inverse Relative Distance Weighted Median Filtering with Edge Enhancement

针对现有的椒盐噪声去除方法未能有效恢复图像的纹理结构和边缘细节,提出了一种带边缘增强的反比相对距离加权均值滤波方法。根据椒盐噪声的灰度特征和无噪像素的分布特征对噪声像素进行识别,再用反比于相对空间距离的加权系数对邻域的无噪像素进行加权,取加权均值作为当前噪声像素的灰度估测值;用拉普拉斯算子对去噪图像进行边缘提取,将边缘图像与去噪图像进行线性叠加,以得到边缘增强的去噪图像。实验结果证明,所提方法具有更好的去噪效果,纹理结构和边缘细节恢复的更好。

To address the issue that effective restoration of the texture structures and edge details could not be achieved effectively by existing salt and pepper noise removal methods,a method of inverse relative distance weighted median filtering with edge enhancement is proposed.Noisy pixels are identified by the intensity characteristic of the salt and pepper noise and the distribution characteristic of noise free pixel,then the noise free pixels in the neighborhood are weighted by the weighting coefficient inversely proportional to the relative spatial distance,and the weighted mean is taken as the estimated value of the processed noisy pixel.The Laplace operator is used to extract the edge of the denoised image,and the edge image and denoised image are superimposed linearly,so as to obtain the edge enhanced denoised image.Experimental results show that the proposed method has better denoising effect and provides better restoration of the texture structures and edge details.