Abstract:

The fast movements of objects and the relative movements between the objects and a camera during the process of image acquisition is one of the main reasons that cause the degradation of images, which is known as motion blur.The blur length and the blur direction are main factors affecting the blur.A novel error parameter estimation method is proposed to identify the blur length and the impact of different restoration methods on the results are compared.Fourier decomposition and Hough transform are adopted to estimate the motion blur direction from the spectrum properties of a blurred image.Experiments validate the good effectiveness of our method.

Key words: motion blur;blur length;blur direction;error parameter curve;spectrum properties;Hough transform