Noise suppression methods in beam hardening correction for X-ray computed tomography

The formation principle of X-ray Computed Tomography (CT) beam hardening and the effect law of noise on the correction result were studied. The suppression methods of noise impact in the process of beam hardening correction were presented from the following three aspects: First of all, aiming at the adverse factors of original beam hardening data, such as noise and measurement error, a method of simplifying beam hardening data was proposed based on histogram statistics of traversing lengths, which removing noise while ensuring the data density uniformity; Secondly, according to the basic characteristics of hardening curve, a new fitting function with good stability of curve shape was constructed, further enhancing the anti-noise capability of hardening model; Finally, an improved calculating method of beam hardening correction was proposed to primarily eliminate cupping artifacts and at the same time, to control the image noises at the original level. CT simulation result shows that these methods are well stable in the computing process of beam hardening correction for CT images contained serious noises, and the correction result is good.