Research On The Motion Image Deblurring Algorithm

Motion blur images produced in camera shake or the object move.The image deblurring is restored from the blurred image to the clear image.If the fuzzy core is unknown or the image has noisy,motion blurred image is difficult to deblur.This article introduces the theory of image deblurring,and then the deblurring of the simulated image and filmed image is studied.The work of this paper is mainly in the following aspects:(1)The improved adaptive median filtering method is studied.In order to reduce the loss of the details,this method only for suspected noise is denoised better.(2)An improved fuzzy kernel estimation method of uniform linear motion is researched.In this method,base on secondary Fourier transform and Hough transform to estimate angle is researched.And then,an improved adaptive of the fuzzy length methods is designed.When the fuzzy length by the autocorrelation estimate is less than the setting threshold,spectrum projection is used.Finally,the fuzzy core of accurately estimate is studied.(3)An adaptive wiener filtering based on the genetic algorithm is studied.This method use the genetic algorithm to optimize the K value of the Wiener filter.This method is not only suitable for the simulation image deblurring,but also has certain effect on the real image deblurring.(4)An improved adaptive median filtering method for RL image deblurring is studied.For the noised image,firstly,improve the noise in the adaptive median filtering,and then use RL algorithm to deblurring.According to the results of computer simulation,the best use of both methods are discussed.(5)A motion image deblurring system based on MATLAB is designed.