Research On Image Reconstruction Algorithm Of Incomplete Projection Data

In CT systems,image reconstruction algorithms usually require a complete projection data set.However,in many practical applications,due to the influence of some objective factors,the image projection data is missing or noisy.As a result,the classic image reconstruction algorithms are no longer satisfactory.Therefore,using incomplete projection data to study image reconstruction algorithms is of far-reaching significance for further research on image reconstruction technology.At present,the main CT image reconstruction algorithms include analytic reconstruction algorithms and iterative reconstruction algorithms.Image reconstruction of incomplete projection data mainly adopts iterative reconstruction algorithms,including limited angle image reconstruction algorithms and sparse angle image reconstruction algorithms.First,for the limited angle image reconstruction algorithm,this paper mainly studies the filtered back-projection reconstruction algorithm(FBP)and joint algebra reconstruction algorithm(SART)based on Shearlet transform,called FBP-Shearlet algorithm and SART-Shearlet algorithm.The main steps are respectively The FBP algorithm and SART algorithm reconstruct the image results to perform Shearlet positive transformation to obtain Shearlet coefficients,discard the Shearlet coefficients smaller than the set threshold,and then obtain the updated reconstructed image by inverse Shearlet transformation.For SART-Shearlet algorithm,the updated reconstructed image needs to be The result is used as the initial value of the SART iterative reconstruction algorithm and continues to iterate alternately.Secondly,for the sparse angle image reconstruction algorithm,a joint algebra reconstruction algorithm(SART)based on guided filtering is researched,which is called SART-GIF algorithm.The algorithm reconstructs the image results of the SART algorithm,and guides filtering through a box filter with a square window radius to obtain the updated reconstructed image.The updated reconstructed image results are used as the initial value of the SART reconstruction algorithm to continue iterating alternately.Finally,the simulation experiments are used to compare the limited angle image reconstruction algorithm and the sparse angle image reconstruction algorithm proposed in this paper.The experimental results show that,for the limited angle image reconstruction algorithm,the SART-Shearlet algorithm can not only effectively remove the noise in image reconstruction due to limited angle projection data,but also better retain details such as image edges and obtain better visual effects For the sparse-angle image reconstruction algorithm,the SART-GIFalgorithm can reduce the mean square error and improve the image reconstruction quality.