A New Image Denoising Algorithm Using Morphological Diffusion
Hamid Reza Moetamedzadeh, Nazanin Afrasiabi, Mohammad Reza Hairi Yazdi · 2013
In this paper, a new algorithm for multivariate noisy environment is proposed. Anisotropic diffusion has a good performance for gaussian noise, but the response of filter is not satisfactory for nonlinear noise. In fact, the diffusion coefficient is a function of gradient magnitude and for nonlinear noise, ordinary gradient estimation fails. The morphological filters can be used for nonlinear noise reduction. The proposed method uses the morphological operators to preprocess the noisy image and uses the denoised image for gradient estimation. The technique proposed in this method enables the linear and nonlinear noisy environment to be well handled. Based on the experimental results, the proposed method achieves a good visual quality and PSNR in presence of multivariate noise.