Logo of the Physikalisch-Technische Bundesanstalt
Production sequence of Si-spheres and interferometrical determination of the sphere volume

Fast and easy-to-use beam-hardening measurement method for computed tomographic (CT) images


To improve the quality of CT measurements, a number of corrections in the tomographic images are necessary. Examples of such corrections are bad-pixel elimination, flat-field (shading) correction, image rectification, deconvolution of the detector’s point spread function (PSF), scatter elimination, and beam-hardening correction. The latter can be described by a look-up table of the relation of measured grey value and radiographed material thickness.

A simple and fast method and a test object were developed for the measurement and correction of the beam hardening artefact in X-ray computed tomographic images. The test object features a nearly conical attenuation body, designed for a certain working distance so that all radial profiles in the projection image represent a linear increase of the material penetration length. Two test objects were manufactured of stainless steel and aluminum to cover a wider X-ray energy range. Two central pinholess made of bronze were included in the design to facilitate the alignment of the device towards the X-ray source. An additional (check) hole allows a simple check of the correction of point-spread-function (PSF) of the detector.

In addition to the "look-up table" for correcting the so-called beam hardening artifact, the thickness of the scintillation layer of the X-ray detector can also be estimated with the help of a simulation. For this purpose, preliminary investigations were performed based on real and simulated data, varying the scintillator thickness in the simulations and comparing the simulation results with real data, see Fig. 2. The beam-hardening curves adapt asymptotically to the simulation curves for a 400 µm thick scintillator layer for attenuation factors up to 10 (see the ordinates in Fig 2). For attenuation factors greater than 10, the measured beam hardening curves deviate from the simulations at the same scintillator thickness. This divergences can be explainded by scattered radiation from the CT cabin and the base plate from PVC, which could not be simulated.

Cone-like test object
Fig. 1. Test device for beam-hardening measurement. (a) Schematic profile highlighting the device’s features and (b) photograph of the manufactured test device made of aluminium

Four measurement results obtained with the cone-like test object
Fig. 2. Comparisons between the real and simulated beam hardening curves with different specimen materials and tube voltages.





Physikalisch-Technische Bundesanstalt
Bundesallee 100
38116 Braunschweig