Establishment of pre-processing station for denoising NOAA satellite images using Legendre Fenchel transformation method

摘要

The main focus of this paper lies in establishing a preprocessing (particularly for denoising) station for National Oceanic and Atmospheric Administration (NOAA) satellite images. Main theme behind this establishment is to make use of new algorithms for the purpose of denoising satellite's images. This paper makes use of total variation method based denoising and Legendre Fenchel transformation (ROF) based denoising for the removal of unwanted pixel information from an image. Further, this paper identifies that Legendre Fenchel ROF model suits better in denoising. The reason behind this identification is because of its computational speed while denoising an image without deteriorating the quality of an image. This paper uses the newly designed turnstile antenna for capturing signals (WXtoImg software is used in converting APT (Automatic Picture Transmission) signal of NOAA to an image) from NOAA satellite. Results are scrutinized and validated through various quality metrics. This station aims at solving the issue of denoising at less process time without degrading the quality of an image by the use of Legendre Fenchel ROF model.