Research On Restoration Algorithm For Global Image Blurring Due To Constant-speed Linear Motion

Restoration of motion image blurring is an important research direction in the field of digital image processing,covering many fields such as remote sensing technology,medical imaging,transportation systems,industrial manufacturing,etc..It has received extensive attention and research.However,the existing motion blur image restoration still has problems that need to be resolved;for example,the point spread function(PSF)estimation is affected by factors such as the cross-bright line in the center of the spectrum,resulting in low accuracy,and obvious boundary ringing occurs after the image is deconvolved.Aiming at the above problems,this thesis has done the following work in the restoration of motion blurred images:(1)Estimating the point spread function of a motion blurred image mainly involves two physical quantities,the blur angle and the blur length.Traditional algorithms are susceptible to interference from the cross-beam in the center of the spectrum and noise in the cepstrum when estimating the motion blur angle.The method adopted in this thesis is to estimate the motion blur angle in the gradient cepstrum image,then set a threshold to suppress the noise in the cepstrum image,and finally obtain the motion blur angle through the phase-consistent edge detection algorithm and the improved Radon transform algorithm.After obtaining the blur angle,rotate the blur image clockwise by the same angle to obtain the horizontal motion blur image,and then use the differential autocorrelation algorithm to estimate the blur length.This algorithm successfully solved the problem of noise interference in the cross-bright line in the center of the spectrum and the cepstrum.(2)Deconvolution of motion-blurred images.In view of the large error of the k value in the traditional Wiener filtering algorithm,this thesis uses the quadratic Wiener filtering algorithm to determine the k value;in view of the problem of the traditional Wiener filtering deconvolution method,which has the problem of the boundary ringing of the restored image,the boundary-based expansion is adopted.Finally,the boundary expansion algorithm is integrated into the quadratic Wiener filtering algorithm,and a quadratic Wiener filtering algorithm based on boundary expansion is proposed to restore the motion blur image.This algorithm eliminates boundary ringing while improving the accuracy of k value.(3)Using the above algorithm,a restoration experiment is performed on the real motion blurred image with Matlab platform.The experimental results show that the algorithm in this thesis can effectively restore the motion blurred image,by accurately estimating the point spread function of the motion blurred image,better estimating the k-value of the Wiener filtering and suppressing the boundary ringing effect.The research in this thesis can solve the problem of recovering the actual motion-blurred image,and has important significance for the research of motion-blurred image restoration technology.