Lei Niu MSc Dissertation 2014/15

Imgae Processing

Supervised by J.Winkler

Abstract

  In rencent years, Image restoration have occupied a crucial part in numerous image system fields, because this process could provide high-quality images under the severe conditions such as various blurs, noise and physical drawbacks.

  This paper is mainly focused on the process of image deblurring, which is to recover the clear image from the blurry image. The work of this paper is first to recover the sharp imgae by a modified algorithm, and then do some further processes, including calculate the error between original image and restored image. And then find out the relationship among three parameters, the size of blur kernel, the number of iteration and the error. Finally, different image deblurring methods are applied to compare their performance through the error surfaces.

Key Words: Image Restoration, Image Deblurring, Blind Algorithm, Error Surfaces