Speckle Noise Removal Using Dwt and FFT with Mask Technique

Abstract

Speckle type noise is a kind of granular-based pattern signal.  It is a special type of noise that is mostly available in images captured by satellites. Removing speckle type noise is the major issue and least considered problem. These kinds of images are taken by an exceptional radar type called SAR (Synthetic Aperture Radar). Speckle-type noise produces an unwanted result. Tactlessly, the occurrences of speckle kind noises disturb edge values and limit the contrast resolution. Differentiating speckle noise from the actual information is very hard due to its behavior. So, it is essential to avoid these issues by decreasing the noise level in both medical images and the images collected from the satellite. The main objective of this research work is to develop two various algorithms based on DWT and FFT with the MASK technique to remove speckle noise with the aid of guided and frost filters. Mainly this algorithm is used to remove speckle noise from satellite images and in an ultrasound image. The performance of the proposed algorithm is assessed based on various metrics such as PSNR (Peak Signal-to-Noise Ratio), mean square error, and structural similarity.