Research On Reconstruction Algorithm For Rotating CL

Computed Laminography(CL)is a technique to reconstruct the interior of an object by using the projection information irradiated from different angles in the thickness direction of the object.Compared with computed tomography(CT),it solves the problems of limited scanning space and serious ray attenuation in plate imaging,and is suitable for non-destructive testing of plate-like objects such as plate-like fossils,printed circuit boards(PCB)and integrated chips.Although CL scan solves the problem that traditional CT is not suitable for plate object detection,the projection data obtained by CL scan does not meet the Tuy complete criterion,and there are edge artifacts and detail blur in the reconstructed image.In addition,because of the characteristics of the CL scanning system,the mutual interference between layers will lead to inter-slices aliasing artifacts,which is shown as "adhesion" between adjacent layers.In recent years,sparse view imaging is also a research hotspot,which can effectively reduce the time and acquisition cost,but its reconstruction quality is poor.Therefore,it is of great practical significance to study the CL image reconstruction algorithm and improve the reconstruction quality under sparse and full-view CL scanning.In this paper,the image reconstruction algorithms for rotating CL are studied,focusing on the total variation(TV)regularization term,prior image(PI)regularization term and optimization framework.The CL image reconstruction algorithms are analyzed and improved algorithms are proposed.The following are the main contents of the study:(1)We propose an iterative reconstruction algorithm constrained by a truncated adaptive weight total variation(TAw TV).The image gradient amplitude is first truncated according to a threshold,and then we design a cosine nonlinear function of truncated gradient amplitude to adjust the truncated gradient adaptively thus the truncated adaptive-weight total variation can overcome over-smoothing when penalizing larger gradient amplitude and isotropic property.Experiments on both simulated three-dimensional(3-D)printed circuit board and 3-D SheppLogan phantom show that the proposed algorithm has noticeable results in artifact suppression and edge-preserving.(2)We proposed a 3-D CL method based on prior images and TV for sparse-angle reconstruction,which is divided into two sub-problems corresponding to two solution steps: the novel Prior-Image-Based Recosntruction(PIBR)step and the TV-based denoising step.In the first step,we devise the final prior image and its corresponding prior mask to ensure the effectiveness of the PIBR.The final prior image is obtained based on two prior images,which are reconstructed from two different oblique angle projections.The prior mask is a binary mask,which is used to record the position of effective information of the prior image.Then the denoising step is based on the TV minimization method using the Split-Bregman(SB)frame.Soft shrinkage operations and Fast Fourier Transform(FFT)are used to efficiently implement the SB frame.Therefore,the algorithm is called PIBR-SB.The experimental results prove the effectiveness of the PIBR-SB algorithm in terms of preserving edges,suppressing inter-slices aliasing and denoising.Through the above methods,this paper has completed the research on CL of PCB,designed the optimization model and its solution algorithm under full-angle and sparse-angle projection,and systematically evaluated its reconstruction performance.The research results of this paper have certain theoretical significance and practical value for promoting the application of CL reconstruction.