Optimal conditions for X-ray imaging by mathematical simulation
1999 (English)In: AIP Conference Proceedings 509 / [ed] Sarah Kallsen, Connie Nessa, Donald O. Thompson, Dale E. Chimenti, Linda Poore, 1999, 665-672 p.Conference paper (Refereed)
Image quality strongly affects the detectability, which is the possibility to detect defects. To obtain maximum detectability it is necessary to conduct the testing with optimal equipment settings. In radiography and computerized tomography based on conventional poly-energetic x-rays, the optimal equipment settings depend on the imaging task. An imaging task is defined as testing of a specific object with a specific defect, defined by composition and geometry. However, the optimal equipment settings (e.g., x-ray tube potential, x-ray filtration and exposure time) are tedious to find experimentally. This is particularly true for industrial applications due to the wide range of imaging tasks. In this work, mathematical models of the image collection process for radiography and computerized tomography have been developed. The objective has been to develop techniques to aid the imaging operator to find optimal imaging parameters. With the models it is possible to find the optimal settings and to predict the detectability of defects in terms of its size as a function of imaged object diameter and these are formulated in terms of detectable detail—object diameter diagram. It is shown that the image quality is very sensitive with respect to the settings, with e.g., slightly non-optimal choice of x-ray filter thickness leading to a loss of image quality that cannot be compensated by varying the x-ray tube potential. Furthermore, non-optimal conditions are found to considerably reduce the detectability of defects, especially for large objects.
Place, publisher, year, edition, pages
1999. 665-672 p.
nondestructive testing, X-ray imaging, radiography, computerised tomography, modelling, digital simulation, optical transfer function, optimisation, random noise, image processing
Engineering and Technology
IdentifiersURN: urn:nbn:se:liu:diva-30136DOI: 10.1063/1.1306113Local ID: 15616ISBN: 1-56396-930-0OAI: oai:DiVA.org:liu-30136DiVA: diva2:250957
Review of progress in quantative nondestructive evaluation, 25-30 July, Montreal, Canada