Certain investigations for removal of noises in digital images and implementation using FPGA

Abstract

Image processing forms a core area in the field of engineering and technology. It involves processing of an image in order to extract the useful information from it. It is widely used in many fields such as medical imaging, scanning techniques, printing skills, license plate recognition, face recognition and so on. The applications of image processing include remote sensing, biomedical imaging applications, automatic visual inspection system, moving object tracking, surveillance etc. A digital image is a two dimensional signal. Images are often corrupted by noises during image acquisition and transmission. The performances of the image processing techniques are affected by corrupted noise. It deals to change in pixel values of the image that do not reflect the true intensities of the real image. Mathematically there are two basic models of noise additive and multiplicative. Additive noise can be modelled easily whereas multiplicative noise is image dependent, complex to model and hence it is difficult to reduce. The additive noise includes impulse noise which is generally two types: Fixed Valued Impulse Noise (FVIN) and Random Valued Impulse Noise (RVIN). FVIN is termed as salt and pepper noise which can be either 0 or 255 whereas in RVIN, the noise is random in nature and have any value between 0 and 255. The multiplicative noise is a speckle noise which is a granular noise that degrades the quality of the Synthetic Aperture Radar (SAR) image and medical images. This is modelled by the random multiplications of the pixel value of the image. Speckle noise in SAR image causes difficulties in image interpretation. It is caused by back scattered signals from multiple targets. newline