Comparison between beam-stop and beam-hole array scatter correction techniques for industrial X-ray cone-beam CT

Abstract In industrial X-ray cone-beam computed tomography, the inspection of large-scale samples is important because of increasing demands on their quality and long-term mechanical resilience. Large-scale samples, for example made of aluminum or iron, are strongly scattering X-rays. Scattered radiation leads to artifacts such as cupping, streaks, and a reduction in contrast in the reconstructed CT-volume. We propose a scatter correction method based on sampling primary signals by employing a beam-hole array (BHA). In this indirect method, a scatter estimate is calculated by subtraction of the sampled primary signal from the total signal, the latter taken from an image where the BHA is absent. This technique is considered complementary to the better known beam-stop array (BSA) method. The two scatter estimation methods are compared here with respect to geometric effects, scatter-to-total ratio and practicability. Scatter estimation with the BHA method yields more accurate scatter estimates in off-centered regions, and a lower scatter-to-total ratio in critical image regions where the primary signal is very low. Scatter correction with the proposed BHA method is then applied to a ceramic specimen from power generation technologies. In the reconstructed CT volume, cupping almost completely vanishes and contrast is enhanced significantly.