Cone beam computed laminography based on adaptive-weighted dynamic-adjusted relative total variation

Abstract

X-ray computed laminography (CL) is an advanced non-destructive testing technique. It allows 3D imaging of the interior of a plate–shell object using projections at a tilted angle, making up for the shortcomings of traditional computed tomography (CT) in imaging plate–shell objects. Nevertheless, the incomplete projections of the cone beam CL result in inter-slice aliasing and cone angle artifacts in the reconstructed image. To address this issue, we present an adaptive-weighted dynamic-adjusted relative total variation (AwDaRTV) minimization algorithm. Based on relative total variation (RTV), the proposed AwDaRTV considers the edge property among adjacent image voxels and introduces weights that are adaptively adjusted with local image gradients. In addition, a dynamically adjusted weight is designed for the regularization term in the hierarchical direction of the reconstructed image. And the weight updates with the number of iterations. To demonstrate the superiority of the proposed algorithm, experiments are carried out on the PCB phantom and the workpiece phantom. Qualitative and quantitative analyzes show that the proposed algorithm works well in reconstructing critical structural features and is effective at suppressing noise, as well as inter-slice aliasing and blurring.