Statistical Quality Analysis of Wavelet Based SAR Images in Despeckling Process

Abstract

Synthetic aperture radar (SAR) images are mainly denoised by multiplicative speckle noise, which is due to the consistent behavior of scattering phenomenon known as speckle noise. This paper presents the basic concept, role and importance of Discrete Wavelet Transform (DWT) in the field of despeckling SAR images and also offers a study of SAR image quality on applying DWT on the speckled image and log transformed speckled image. Log transform operation plays a decisive and comfortable role in despeckling SAR images as this operation changes the multiplicative behavior of the speckle noise to an additive which enables to use the additive noise restoration model efficiently. Wavelet transform has now become important in the field of image restoration although being in practice for a decade. Wavelet transform allows both time and frequency analysis simultaneously around a particular time. This transform is most appropriate for the non-stationary signals, so it deals with satellite imagery in a more efficient manner. The major part of this paper is revolving around DWT image decomposition with its role and practical implementation on the speckled image and log transformed speckled image. All the experimental results are performed on the SAR images.