An optimized filter design approach for enhancing imaging quality in industrial linear accelerator

Area of Science:

• Materials Science
• Nuclear Engineering
• Medical Imaging Physics

Background:

• Polychromatic X-rays from linear accelerators (Linacs) cause hardening artifacts, hindering accurate defect identification.
• Current empirical filter methods lack robust scientific metrics for optimization.
• Effective defect detection is crucial for safety in nuclear reactor components.

Purpose of the Study:

• To develop a scientifically-based filter design method for optimizing X-ray computed tomography (CT) imaging.
• To balance ray intensity and energy impacts on image quality for artifact reduction.
• To enhance defect visualization in graphite materials for high-temperature gas-cooled reactors (HTRs).

Main Methods:

• Utilized GEometry ANd Tracking (Geant4) simulations to obtain X-ray spectra for various filter materials and thicknesses.
• Generated reconstructed images using simulated spectra and incident photon counts.
• Evaluated image quality by comparing intensity differences, noise levels, and hardening indicators in defective and non-defective regions.

Main Results:

• An optimized filter was designed and applied to a Linac-based X-ray CT system for HTR graphite defect detection (2μm defects).
• The filter reduced X-ray hardening effects by 22%.
• Achieved a Defect Contrast Index (DCI) of 3.226, indicating significantly improved defect visibility.

Conclusions:

• The developed filter design method effectively mitigates hardening artifacts in X-ray CT images.
• Optimized filters based on Average Difference (AD) and DCI parameters enhance defect image quality.
• This approach provides a scientifically sound basis for filter selection in critical industrial inspections.