Wavelet Transform Technique Applied to Satellite Image Denoising

Abstract

Satellite images either digital or analog must have certain elements that are accidentally introduced during the processing of capturing as a result of weather or system sensor known as electronic noise. However, several attempts and advances have been made by academicians, industries and intelligent security agencies to remove this noise. It has been a nagging problem in the area of computer vision, image processing and artificial intelligence to denoise satellite images and noise removal is among the significant components in satellite image analysis. The aim of this research work was to denoise the satellite image of Sambisa forest using the wavelet transform technique. Satellite images of Sambisa forest captured by Landsat satellite in 2007, 2013, 2014, 2019 and 2021 respectively with their associated Geo-referenced 11.2503° N Longitude and 13.4167° E Latitude were downloaded from the United States Geological Survey (USGS) website. The images are acquired as Zipped Geo-referenced Tagged Image File Format (GeoTIFF). Color Composite bands of natural colors (bands 2, 3 and 4) are combined using the ArcGIS software and RGB image were obtained. Wavelet transforms denoising technique was used to filter noise from the images, which was implemented using the wdenoise2() function in MATLAB 2021.