Noise suppression in beam hardening correction of X-ray industrial CT

In X-ray industrial Computed Tomography (CT), when Conventional Linearization (CL) correction method is used to eliminate beam hardening effect, it enlarges the projection noise and thus the signal-to-noise ratio (SNR) of CT image is reduced. To overcome this shortcoming, an improved linearization (IL) correction method is proposed. Based on the character of projection noise, an adaptive filter is designed to extract the noise from the original projection data, and employs the noise to evaluate the correction errors. According to the errors, the correction results are revised to suppress the projection noise enlargement. IL correction method is examined by different source voltages and tested materials. Experimental results show IL correction method can eliminate beam hardening effect and suppress the projection noise enlargement effectively in correction process. Compared with CL correction method, IL correction improves the SNR of CT image about 6 times.