An Image Denosing Algorithm for Strong Noise Image Based on Structural Similarity of Contourlet Domain via Grey Relational Analysis

摘要

Aiming at solving the problem of image denoising in strong noise image environment, this paper proposes a new algorithm. First of all, the noisy image is transformed into Contourlet domain via multiscale decomposition to obtain frequency domain information in different scales and directions. Then, according to the distribution and attenuation characteristics of edge and noise in different frequency bands, edge and noise can be distinguished by grey correlation degree, thus the image quality is improved by edge enhancement and noise suppression. The simulation results show that the new algorithm can obtain higher peak signal-to-noise ratio (PSNR) and improve the visual effect of the image compared with more than ten methods such as spatial filtering, wavelet adaptive threshold denoising and Contourlet adaptive threshold denoising, etc.