Novel Method to Assess Motion Blur Kernel Parameters and Comparative Study of Restoration Techniques Using Different Image Layouts

Abstract: In general, images are yielded to preserve or represent convenient information. The recorded image can perpetually characterizes a degraded version of the original image caused by inadequacy in the imaging and capturing process. There exists several kinds of degradations that need to be considered to restore degraded images. Such kinds of degradations are blur, illumination and color imperfections, geometrical degradations and noise. Among those, two major categories of degradations namely, motion blur and noise are considered in this study. Motion blur occurs when there is movement of element or camera or both during exposure time. Besides to this blurring effect, noise also debase any recorded image. Therefore, restoration of motion blur image and reduction of noise are very important in many circumstances like identification of criminals and tracking vehicles number plate. For restoring degraded image, it is important to know the degradation function, commonly point spread function (PSF) of blurred image. The point spread function is defined by two parameters called, blur direction and blur length. This study focuses an approach to assess the parameters from blurred and noisy images. Furthermore, this study presents comparative analysis of di\fferent non-blind image restoration techniques based on Wiener and Lucy-Richardson algorithm for different types of image layouts such as .tif (Tag Index Format), .jpg (Joint Photographic Experts Group) and .png (Portable Network Graphics). The experimental result illustrates blur images without noise and with noise are successfully restored.