Image Reconstruction Algorithm Of Incomplete Projection Data

Traditional computed tomography(CT)utilizes different attenuation coefficients in different density distribution of X-ray to get the image of object.Due to its advantages in lossless,high precision and visualization,CT is widely used in medical imaging,industrial nondestructive testing,safety inspection and so on.The projection data and reconstruction algorithm are two key aspects of CT.High-quality reconstruction image can be obtained by employing the analytic reconstruction algorithm under the condition of complete projection data.However,limited to environment and imaged objects in the practical applications,there are often incomplete projection problems.For example,sparse incomplete dataset can be collected to reduce the dose of X-ray radiation.Indeed,achieving high-quality reconstructed image from incomplete dataset by use of novel image reconstruction algorithm is of great academic and application value.In this paper,the image reconstruction algorithms from incomplete projection caused by sparse and limited-angle problems have been studied.We focus on the following two aspects: i)research and improve the weight projected system matrix for iterative reconstruction algorithm;ii)further investigate regularization optimization iterative reconstruction algorithm.The main contents can be summarized as follows:(1)The weight coefficient matrix has an great impact on the image quality by utilizing iterative reconstruction algorithm.Firstly,the RayBox Intersection weight matrix algorithm and the logic loopholes have been deeply analyzed.To avoid the disadvantages of RayBox weight matrix algorithm,an improved algorithm is adopted to make up for logic loopholes.In addition,the improved algorithm and the classical bilinear interpolation algorithm have been employed to verify the effectiveness of the improved RayBox Intersection algorithm.(2)The problem of sparse incomplete projection is researched.For reconstructed image resolution is too low from sparse incomplete projection,the improved RayBox Intersection algorithm was introduced into the classical ASDPOCS regularization reconstruction algorithm.As the subpixel has the property of improving the image resolution,ASDPOCS algorithm based on subpixel is proposed.Qualitative and quantitative experiments demonstrate that the algorithm can effectively improve the reconstructed image resolution.(3)The problem of incomplete projection due to limited angle is studied.As for limited-angle problem,the augmented matrix has been constructed by using the similarity of reconstructed image structures for different iterations,and the singular value threshold algorithm is applied to denoise under the constraint of a prior image.This method has some inhibitory effects on the noise of reconstructed image with limited-angle problem.However,the edge of the reconstructed image needs to be further improved.In order to further improve the image quality and the convergence of the algorithm,the algorithm is improved.In other words,the edge of the image has been extracted and integrated into the augmented matrix.The experimental results show the above two reconstruction algorithms can further improve image quality with limited angle.