Research On Image Deblurring Based On Wiener Filter With Subjective Evaluation

Information stored in images plays an irreplaceable role in all fields of society.However,in the process of image acquisition,it is always affected by the outside world or due to its own reasons,so that the image quality will appear blur.In order to make the information in the image clear and restore the original high-definition quality image,how to use the blurred image to restore the clear image has become a hot research at home and abroad.As a branch of image processing,image deblurring lays an important foundation for subsequent pattern recognition,so it has great research value and significance.Image deblurring refers to the process of reconstructing degraded image to restore its original information to the maximum extent.In recent years,with the continuous development of all kinds of shooting machines,how to de blur the image in time after being blurred has become a research hotspot.I draw on the research experience of predecessors,extract the essence,and propose a new method of image deblurring.First,we propose a function of image fuzzy judgment to judge whether there is fuzzy phenomenon in images.Secondly,we divide the image into image blocks and process them on the basis of image blocks.Finally,we combine the principle of minimum variance of Wiener filtering on the brightness and color modes of images.The main contents of this paper are as follows.1)Research on image quality evaluationThe effect of traditional image de blurring is based on subjective evaluation and objective evaluation.Objective evaluation is calculated by specific mathematical formula,and the common ones are mean square error,signal-to-noise ratio and peak signal-to-noise ratio.Because there is a large error between the blurred image and the original image,the mathematical model is against the traditional principle of "the smaller the mean square error is,the better" and "the larger the peak signal-to-noise ratio is,the better".To solve this problem,this paper proposes that the evaluation after de blur should be based on the subjective evaluation.In view of the possible errors in the subjective evaluation,this paper designs a new algorithm of image quality score after de blur.2)Research on image deblur model.Classical inverse filtering,Wiener filtering and R-L deblurring all have some disadvantages.In view of these disadvantages,this paper proposes a new method of deblurring based on the previous research results and the method of Wiener filter deblurring,that is,combining the RGB channel of the image to carry out the fuzzy kernel estimation respectively,and combining the ratio of the original image and noise of Wiener filter to realize the adaptive deblurring.Experimental results show that the method proposed in this paper can reduce the ringing phenomenon of images to a certain extent.3)Judgment of image bluringIn order to detect whether there is blur in the image,this paper proposes a function to judge whether there is blur in the image.Considering the gray level image information and image map features,this paper proposes a method to judge image gradient function based on singular value,which enhances the efficiency of image de blur.Based on the image in the fuzzy image database,the experimental results show that the function of judging image gradient based on singular value can improve the efficiency of image deblur.Based on the adaptive Wiener filter image de fuzzy research,and in the fuzzy video frame mountain,image data and image block de fuzzy,and achieved good research results.In the future research,this paper will consider adding a variety of fuzzy image deblurring,and further open the research based on image deblurring.