A Novel Beam Hardening Correction Method for Computed Tomography

As a polychromatic x-ray beam passes through matter, the lower energy photons are preferentially absorbed mainly because of photoelectric absorption. So the spectrum becomes narrow. This leads to various artifacts in reconstruction image and imaging degradation. Cupping artifact is one of them. A novel beam hardening correction method based on reprojection is presented here. Firstly, the original CT (computed tomography) image is segmented into a binary image. Secondly, reproject the binary image for getting the path length information of x-ray penetrating specimen. Thirdly, lit the relationship between the thickness of the specimen and the original polychromatic projection by polynomial function to gain the beam hardening correction model. Then, according to the model, the original polychromatic projections are corrected. CT image reconstructed from the corrected projections shows that the cupping artifacts are effectively eliminated. However, new noise arise in the CT image .To improve the result, one adaptive filtering method are described to denoise the original projections. The final excellent result indicates that the method is effective. Compare with conventional method with the polynomial fit technique, this method doesn´t need wedge-shaped phantom that is used to acquire the correction model. And the noise of the CT image is lower than the conventional method.