The Research Of Motion Deblurring Algorithm

This dissertation presents a new algorithm for removing motion blur from a singleimage. Restoring blurred image，as one of the hot issues in the field of image processing，has important significance in improving the image quality. Motion deblurring has widelybeen used in the field of astronomy, military, roads, transportation, medical imaging,industrial control and manufacturer, and etc. In recent years, a variety of methods forremoving motion blur of an image have been proposed, but most of the algorithms are toocomplex to preserve image details and eliminate various ringing artifacts at the same time.This dissertation first describes the fundamentals of motion-blurred images, and thesignificance of the study on motion blur image restoration. A new method is then presentedwhich is more efficient than the existing motion deblurring algorithms. TheLucy-Richardson algorithm uses inversion formula to recover motion blur images, whichresults in good restoration effect in the smooth region, but poor recovery in thehigh-frequency region. Therefore one needs to combine the LR algorithm with otheralgorithms to recover motion-blurred images. For this characteristic, we propose to improvethe algorithm, the new algorithm uses Gaussian filter suppress high frequency region toprotect the edge of original image. Using an adaptive function, high-frequency part of themotion-blurred images and smooth region can be distinguished; different methods arecoped with different areas. Ultimately, the image which was restored by the new algorithmcan not only ensure the edge of the image to get recovered, but also eliminate variousartifacts.