The Research Of Motion Deblurring Algorithm

Motion blur widely exists in daily life,bring inconvenience for the image information getting. Making an in-depth research on motion blur, find an effective way for image restoration by combining method of mathematical analysis, algorithm design and experiment simulation. According to the mechanism of motion blur produced, Propose solutions for different situations motion blur. For getting the interested information from the blur image effectively and steady, based on classic Lucy-Richardson algorithm, new image restoration algorithm to be proposed for the motion blur restoration and the partial- motion blur restoration. Make improvements of classic Lucy-Richardson algorithm to improve the ringing and speed.By using the edge information of image to get the point spread function, an image restoration algorithm is proposed for the global motion blur problem. A single blur image is scaled to a series of similar images and using each image information to get the point spread function by itself, and under the control of certain conditions,achieve an accurate estimate of the point spread function and image restoration. The result of the experiment shows that The global motion blur restoration algorithm has a strong ability to resist noise and increase processing speed.Against partial motion blur problem, a restoration algorithm is proposed by combination of alpha channels and image segmentation. Firstly, based on blurred image information and the relationship between alpha channels and point speed function get the parameters, estimate the point speed function and achieve the motion restoration. Then,taking advantage of the knowledge of image segmentation, separate the background image and foreground image. Finally, obtain a restored image according to the image synthesis method. The result of the experiment shows that the partial-motion blur restoration algorithm can accurately estimate the blur angle and length, can estimate the point spread function exactly and rebuild image, and the image synthesis method is effective to solve the problem of blurred image restore.