Answers are grounded in this paper and cite their sources.
In this paper, a new beam hardening correction (BHC) method for middle-energy industrial computerized tomography (CT) is presented. Our method is derived from linearization and is straightforward without iteration involved. The linearization is commonly used as a preprocessing method in BHC. Conventionally, only one-material objects can be conveniently corrected by linearization. In industrial CT, two-material objects, especially cylinders with high-Z material outside and low-Z material inside are frequently encountered. Our approach focuses on this kind of objects. The new method works well as long as the two-material object meets the conditions that the thickness of the outer material (usually wall) is thick enough and the second-order item of the Taylor expansion of the linearization is relatively small. We pointed out and proved that there is an approximately constant scaling factor difference between our linearization step and an ideal correction based on prior knowledge of objects. The scaling factor magnifies the attenuation coefficient of the inner material after reconstruction. Therefore, a weighting function is introduced into our algorithm as a restoration. To sum up, there are three steps in our method: 1) correct raw projections by the mapping function of the outer material; 2) reconstruct the cross-section image from the modified projections; 3) scale the image by a weighting function. With this method, the beam hardening artifacts are greatly reduced and the overall attenuation coefficients are accurately obtained. We also presented a compensation step to remove the countercupping artifacts in case that the conditions are not fully met. Our method is well verified in both numerical simulations and practical experiments on a 450-KeV CT system.
No takes yet. Share an insight, caveat, or question.
Gao et al. (2006) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: