Blur Parametric Estimation on Natural Images for Linear Motion, Out-Of-Focus and Gaussian Blur for Blind Restoration

Abstract: Image restoration and recognition in blurred and poorly illuminated images is difficult. The recognition and the restoration factors are of vital importance in this endeavour. Image blur is difficult to avoid in many situations and can often ruin a photograph. Image deblurring is a process, which is used to make pictures sharp and useful by using mathematical model. The blind image restoration used to remove the blurs in the image. In this, the modified radon transform used to handle the three types of blurs in the given input image. The proposed method is used for estimating the parameters of linear motion, out-of-focus and gaussian blurs. The method is implemented by analyzing the blurred images spectrum. The modification in the radon transform is proposed in this. Blur parameters are identified using these modifications of radon transform i.e. radon-d, radon-c transform. By fitting an third order polynomial function that accounts separately for the image spectrum and the blur frequency blur parameters are estimated.

Keywords: Gaussian blur, linear motion blur, Image restoration, Out-of-focus, Radon transform, Radon-d, Radon-c, Spectrum of blurred images.