Orthogonal Compactly Supported Near-Symmetric Wavelets in Denoising Satellite Images

Objectives: This is to investigate the performance of two orthogonal, compactly supported, near symmetric wavelet families’ viz., coiflets and symlets in denoising satellite images. Methods/Statistical Analysis: We use the Stationary Wavelet Transform (SWT) which is a shift-invariant transform and three levels of image decomposition. Additive White Gaussian Noise (AWGN) is used to produce the corrupted images used for the study. Findings: The study identified the most suitable wavelet in each of the above two wavelet families, for denoising satellite images. A comparison between the denoising performances of these wavelet families has also been made. It has been found that the denoising performance of wavelets belonging to both the families decrease as the wavelet order increases. Applications/Improvements: The study finds application in the denoising of satellite images and images obtained by space exploration and such images corrupted by a combination of several types of noise distributions.