A discrete grayscale prior-based exterior reconstruction algorithm for polychromatic X-ray CT

Abstract

Exterior CT imaging is a special X-ray imaging problem that allows for nondestructive testing of relatively large tubular samples by using smaller detectors. However, due to the incomplete nature of the exterior projection data, the exterior CT imaging problem is highly challenging. In this study, we introduce a new CT reconstruction model for polychromatic spectrum exterior problems, called the Polychromatic Exterior Discrete Grayscale PAEDS (PE-DG-PAEDS) model. This model is based on the prior of discrete grayscale values in images and introduces a radial regularization term using polychromatic spectrum information for exterior CT reconstruction. Additionally, an alternating minimization method and the Discrete Algebraic Reconstruction Technique (DART) algorithm are used for alternating iterations to provide a solution algorithm for this model. Experiments conducted with both simulated and real data have validated the proposed model and algorithm. The results indicate that the method effectively suppresses artifacts associated with polychromatic X-ray CT exterior problem.