The existence of noise in many applications such as industrial, military and medical images is one of the important factors that affect the quality of the produced images and may disrupt the proper functioning of these applications. Considering the need to perform some of these applications in a real-time manner, providing real-time image noise reduction methods is a necessity. Impulse noise is a type of noise which corrupts images due to the malfunctioning of the acquisition device or electronic equipment. Up to now, different hardware and software-based techniques have been proposed to remove this type of noise. In this thesis, for improving the performance of current approaches, three different hardware-based methods are proposed for random-valued impulse noise removal. In these proposed methods, image pixels are categorized to different groups based on the image structure as well as the effect of noise on image pixels. Then noise detection and reduction steps are performed with the aim of preserving of the structural characteristics of the image. Furthermore, based on the similarity of pixels in their neighborhood and the intensity of noisy pixels, a low complexity hardware-based approach is proposed to remove the salt and pepper noise. Finally, a low complexity method for removal of reflected light in endoscopic images is proposed. Simulation results of the proposed methods using a publicly available dataset showtheir correct functionality. Keywords: 1-Image enhancement 2-Impulse noise 3-Hardware implementation 4-Low complexity 5- Endoscopic images
