Study Of Motion Blurred Image Restoration Algorithm

Image restoration is the process and method to obtain clear images from degradedimages. Which the motion blurred image restoration is an important branch in the imagerestoration technology, and has been widely application value in people’s productionand life in each domains. Motion blur is caused by the relative motion between theimaging system and the targetIn the research of motion blurred image restoration, pointspread function (PSF) estimation and restoration algorithm for motion blurred image arethe hotspots in the research area of digital image processing in recent years.The piont spread function in motion blur image degradation model is an importantparameter, i.e. blur length and blur angle.(1) In order to improve the efficiency of image restoration, used estimate the blurparameter an improved algorithm named quadric transform method (QT) isproposed and image restoration through the analysis of motion-blurred imagerestoration process.(2) We analysed the frequency spectrogram which obtained by using a Fourierhomomorphisms transformationand then getting the initial motion blur image, andon the basis of frequency spectrogram and some spectral characteristics we proposethe spectrum image was applied once Fourier with State transforms, The movementblur direction was estimated through Radon transform.(3) Then the movement blur length along this direction rotate was obtained usingslef-related differential operations. taking movement blur direction and movementblur length as parameters, choosing the appropriate image restoration algorithm torecover the motion blurred imag.By analyzing each principle and each experimental test, study four kinds of imagerestoration methods such as imverse-filter, constrained least square method,Richardson-Lucy and wiener filtering. this study provided the advantages andlimitations for each method. Which wiener filtering method has a certain inhibitioneffect of noise and avoid appearing noise amplification in the inverse-filter excessive.The four restoration to make comparison, choice the method of Wiener filtering forimage restoration of motion blurred images finally. In the estimation of the motion blur point spread function, combination recovery methods for restoration of motion blurredimages, and in the Matlab platform come up simulation test gives experimental resultsand result analysis. The experiment shows that the method is accurate estimation ofpiont spread function. It also has better robustness and accuraty in the noisy images, theideal restoration effect.