Image Restoration from a Single Blurred Photograph
Yang Li · 2016
The blur which is caused by the relative motion between the camera and object scene is called motion-blur. The blur direction and the motion-blurred image are essential for image restoration, which is degraded by the relative motion. And directly affect subsequent image recognition and image analysis. The recently developed approach based on a fractional directional derivative operator with noncausal is derived. In order to have a grasp of image restoration from motion-blurred photograph, a partial derivative operator is deduced on the basis of fractional derivative in this paper. According to use of real and synthesized motion blur images quantitative and qualitative performance evaluations are carried out. Noise is the sensitive factor of this method. The experiments based on a noncausal fractional partial derivative mask verified that this method provides better immunity to noise in comparison and accuracy of identifying motion blur than the method based on integer-order differentiation does in particular large blur length.