Research On Deblurring Algorithm For Long Exposure Images

Images carry information and data.Digital image processing technology plays an important role in promoting human production and living standards.Computer image processing technology has been applied in aerospace,road traffic,biomedicine,industrial detection,communication,security,etc.The phenomenon of blurred images caused by vibration of the imaging system is very common.To solve this problem,this paper will use image processing technology to achieve image deblurring.Since there is only a translation relationship between short-time exposure image sequences,a high signal-to-noise ratio image can be obtained by accumulating multiple short-time exposure images,so as to achieve the purpose of image deblurring.Firstly,this paper introduces the theoretical knowledge of common image registration algorithms.Compared and analyzed the feature extraction algorithm s based on point feature.The result shows that the SURF algorithm has high accuracy,fast speed,and strong robustness.Therefore,the SURF algorithm is selected for feature extraction in this paper.Then the paper gives the design scheme of the deblurring algorithm.The design scheme of the deblurring algorithm in this paper is as follows: In the first step,detect feature points and then the false matching points are removed.In the second step,a window is opened around all matching points,and the matrix-multiplied DFT method is used to estimate the sub-pixel offset in the window,and the average value of all offsets is used as the final offset estimate value.The third step is to realize the fusion of multiple short-time exposure images through interpolation amplification,translation and accumulation.Secondly,this paper proposes a new interpolation algorithm.New interpolation algorithm combines the nearest neighbor interpolation and bicubic interpolation.By designing a reasonable interpolation kernel function,not only can the problem of high gray value change rate between adjacent points in the bicubic interpolation method can be reduced,but also the edge details of the image can be effectively protected and the quality of the image after interpolation can be improved.Thirdly,the SURF algorithm is improved.The traditional SURF algorithm has the problems of too many feature points and false matching.In order to improve the efficiency and the accuracy of matching,this paper uses slope constraints to eliminate the false matching point pairs.By setting the slope threshold reasonably,the false matching point pairs can be effectively eliminated.After the slope constraint removes the obvious false matches,the RANSAC algorithm is used to further eliminate the false matches.Finally,in order to verify the stability and effectiveness of the algorithm,the mathematical software is used to carry out the algorithm simulation experiment according to the deblurring algorithm flow.The results show that the algorithm can improve the fuzzy phenomenon to some extent.