A new non-linear filter for image denoising

Noise, particularly Gaussian noise, can distort images and reduce information. The James Webb Space Telescope (JWST) is vital for understanding the universe and capturing images that inspire people. This paper introduces several nonlinear filters to remove Gaussian noise from the Carina Nebula Image, first image taken by JWST. The filters consider similarities and distances between central pixels and preserve edges as advanced features. However, in this paper, we propose a new nonlinear filter. This filter features a series of non-linear operations starting with cross- validation (CV) techniques to extract the appropriate smoothing parameter of the image. This parameter is used to create the density function, which is then used to apply a total variation noise reduction technique to eliminate the noise. As a result, these candidates are critical to achieving space exploration’s objectives, showcasing technology innovation, and expanding scientific understanding. The results demonstrate that the suggested filter outperforms other noise reduction filters in image restoration at various Gaussian noise densities, yielding PSNR and SSIM values of (52.12) and (0.99), respectively. When the noise ratio is 0.01, the filters’ output surpasses the differences in noise ratio.