Motion-blurred Image Restoration Algorithm And Implementation

Motion blurred image restoration problem is one of most important issuesof image restoration techniques. Motion blur is extremely common in daily life,such as the daily life of photo taken, traffic and filmed in aircraft target imag-es so there will be more or less motion blur problems. Motion blurred imagerestoration techniques can be applied to many fields such as traffic safety, med-ical safety, criminal cases and military fields. Due to human information obtai-ned from outside the great majority are from the image, solve the motion blurbecomes more and more important.In the problem of image restoration, estimation of blurring kernel is the fi-rst and foremost important step. Quality of restored image highly depends upo-n the accuracy of the estimation. This paper focuses on the problem of motio-n-blurred image restoration, according to the characteristic of the motion blur,establish the motion-blurred image degradation model, and proposes a new met-hod of restoring motion-blurred images. First detect the blur angle of the moti-on image: Secondary Fourier transform the motion-blurred image, and then usethe method of bit-plane slicing with Radon transform to estimate uniform moti-on blur angle. Then estimate the blur length: rotate tspectrum of motion-blur i-mage, The direction of the light and dark stripes will be horizontal, estimate t-he distance of the dark stripes, then the Motion blur length will be obtained.af-ter that, the motion-blurred image of the point spread function (PSF) can be d-etermined. Finally, using wiener filtering and inverse filtering for image restor-ation. The experimentalresults show that the proposed method is capable of esti-mating blur parameters more accurately than the existing methods.