A Framework for Performance Assessment of Beam Hardening Correction Algorithms in Industrial Computed Tomography

Industrial computed tomography (CT) images often suffer from beam hardening (BH) artefacts. Since beam hardening correction (BHC) is crucial in industrial CT, it is worth acquiring a good understanding of the BHC algorithms. The purpose of this work concerns the development of a phantom-based framework to assess the performance of BHC algorithms in industrial CT, and its application on several existing algorithms. The assessment is carried out from the perspectives of image quality and dimensional metrology on simulated data sets. In image quality assessment, the quantification of BH artefact is complicated by the presence of noise. A novel method is developed to allow robust BH quantification across a range of noise level and artefact reduction. Assessment results of state-of-the-art BHC algorithms are given.