Research On Key Technology Of Defocus Image Processing

With the development of science and technology,the application of digital image processing technology is more extensive.In the actual work of criminal investigation and physical evidence identification,forensic images are sometimes defocused images due to inaccurate focus of the imaging system.We need to use the defocus image restoration technology to restore them so that we can extract the information we need from the images.For the defocus image restoration technology,this thesis mainly does the following two researches:Firstly,for the defocus blur radius estimation.The traditional differential autocorrelation method uses a laplacian operator,this operator has a large amount of calculation when calculating.In this thesis,based on the shortcomings of traditional algorithms,a defocusing radius estimation method based on horizontal gradient operator is proposed.Experiments show that the method can estimate the defocus blur radius more accurately.Secondly,some traditional defocus image restoration algorithms are introduced.Including: inverse filter restoration algorithm,wiener filter restoration algorithm,etc.In view of the shortcomings of the traditional wiener filtering algorithm,the traditional wiener filtering algorithm is improved.The information entropy function is introduced to perform second wiener filtering on the image.Experiments show that the algorithm can effectively recover the defocused images.Finally,the restoration experiment of defocus blurred image is carried out,and the experimental results are analyzed.The subjective and objective evaluation of the restored images prove that the method in this thesis can effectively restore the defocus blurred images.