Restoration of Images Corrupted by Multipath Fading Channel with Weighted Encoding

Abstract

Multipath fading affects the radio communication links in one form or another. Rayleigh and Rician are the two types of multipath fading channels. During the transmission of data over these channels, the images are affected by many types of noise similar to Additive White Gaussian Noise (AWGN), Impulse Noise (IN) or the combination of both called as “mixed noise”. Removal of such noise is a critical and challenging work. The noise spreading in this case does not have any predefined model and due to this, the quality of the image further reduces. In the proposed method, the mixed noise is removed using Weighted Encoding with Sparse Nonlocal Regularization (WESNR). The Weighted Encoding technique performs better when compared to the existing image denoising methods. The parameters, PSNR and SSIM are considered to compare the performance of Adaptive Median Filter (AMF) and WESNR.