Mapping pixel size and positions of an energy dispersive line detector for X-ray inspection of luggage

Contact:
Jan Kehres, Research Engineer, DTU Physics, jake@fysik.dtu.dk

State of the art X-ray scanners for luggage inspection utilize dual energy radiography to perform material identification. This method yields however a high number of misclassifications due to strong overlap of the objects and vague assumption of the object thickness. The accuracy can be drastically improved by taking projections at different angles and reconstruct the volume X-ray absorption properties. Further improved can be achieved by using spectral detectors to resolve the X-ray absorption coefficient for a number of different energies instead of only for transmitted photons with high and low energy as performed in dual energy radiography.

We are currently investigating the feasibility of using limited angle spectral tomography for extracting the absorption coefficient for several energies to identify prohibited substance in luggage. However, this novel energy resolving detectors currently have a number of issues, i.e. a not uniform pixel sizes due to the not uniform attachment of the contacts to the detector crystal. Furthermore the gaps between the four different crystals of which each detector consists are not exactly known. Neglecting the pixel size and position of the pixel in the reconstruction leads to artefacts in the reconstruction and hence yields lower accuracy during the threat identification.

The NEXMAP section at DTU Physics recently developed a new tomography setup using energy-resolving line detectors. All component of this equipment including motors, X-ray generator and the detector are implemented in a control software, which can be programed with a macro language, allowing to perform experiments with a minimum of required manual work.

Your task will be scanning all 256 detector pixels of two line detector modules through a small beam formed by the X-ray slit installed in front of the X-ray source. You will learn how to operate the setup and how to program macros. The data analysis you will perform with MATLAB® and you will be in charge of to develop the required scripts. Afterwards you will do spectral reconstructions of a data set you acquire of a known phantom. The reconstruction you will perform with and without the exact pixel positions and pixel sizes and compare both reconstructed slices qualitatively. We will assist with the reconstruction and the reconstruction scripts will be provided by us.


A view of the new radiography/tomography setup with energy dispersive line detectors mounted (left) and a photo 3 spectral line detector modules (right).