Restoration of motion blurred images

Degradation of digital images is often caused by camera movement when a long exposition time is set. The blurred images can be enhanced by either iterative or non-iterative methods. From the non-iterative methods, Wiener reconstruction is often used. Iterative methods are slower in nature but usually better results can be achieved. A well known method for blurred image restoration is Richardson and Lucy algorithm. The Richardson and Lucy algorithm often achieves better visual results compared to the Wiener reconstruction, however, it is iterative and slower. As with many iterative methods the problem is how to find such a number of iterations so that the resulting image is of sufficient quality. Generally, it is not possible to tell in advance how many iterations will be needed in order to achieve sufficient image quality. The aim of this article is to present a method that finds the stopping criteria without human interaction. Our iterative algorithm stops when the restored image is of sufficient quality. The Television Study Organization classification was used as a subjective image quality criterion. The experiments were performed on 100 blurred images. The results show that the most of images can be restored with fine quality without a human interaction.