Development of a Technique for Increasing Efficiency of Speckle Noise from Images

I. Introduction An image is defined as artifact or visual representation which is recorded by visual perceptions and then further digitized to convert it in aform which can be stored in the computer memory or some type of storage media such as hard disk.. After the image has been digitized, it can further be operated by various image processing operations. The unwanted material or particles or electric signals degrade the image and they have to be removed for improving its quality. These unwanted particles constitute the noise. Appearance of the noise in the images can be in the form of dots and this is due to the signals getting corrupted by the noise. Removing the noise from the images is defined as denoising. There are different type noises which can get incorporated in the image, such as impulse noise, gaussian noise, additive white gaussian noise, shot noise, quantization noise, anisotropic noise and multiplicative noise. Out of all these, multiplicative noise is most difficult to remove. Multiplicative noise can be defined as unwanted random signal that gets multiplied into some relevant signal during capture, transmission, or other processing. Speckle noise is an example of multiplicative noise which is commonly present in SAR (Synthetic Aperture Radar) images. This paper describes speckle noise, various techniques used in despeckling an image, flaws in image despeckling models which lead to problem formulation. II. Speckle Noise Speckle noise is a granular form of 'noise' that is known for degrading the quality of the active radar and synthetic aperture radar (SAR) images. It is a multiplicative noise, which makes the visual evaluation in ultrasound imaging quite difficult. Speckle noise suppression is a vast field, a lot of people are working on techniques to reduce it without tampering the original image to a great extend. Coherency in signal processing is similar to correlation in statistics as the signals have similar phase and frequency. This coherent processing of backscattered signals coming from multiple distributed targets, results in speckle noise. The edges and fine details of the image are affected by the presence of speckle noise, resulting in limiting the contrast resolution of the image as well. III. Despeckling Denoising is a general term given to the removal of all sorts of noises from an image. For speckle noise removal, despeckling is the technique used specifically. Based upon different mathematical models of the phenomenon several different methods are used to …