An algorithm to unveil the inner structure of objects concealed by beam divergence in radiographic image acquisition systems

Abstract

Abstract. Two main parameters rule the performance of an Image Acquisition System, namely, spatial resolution and contrast. For radiographic systems using cone beam arrangements, the farther the source, the better the resolution, but the contrast would diminish due to the lower statistics. A closer source would yield a higher contrast but it would no longer reproduce the attenuation map of the object, as the incoming beam flux would be reduced by unequal large divergences and attenuation factors. This work proposes a procedure to correct these effects when the object is comprised of a hullor encased in it -possessing a shape capable to be described in analytical geometry terms. Such a description allows the construction of a matrix containing the attenuation factors undergone by the beam from the source until its final destination at each coordinate on the 2D detector. Each matrix element incorporates the attenuation suffered by the beam after its travel through the hull wall, as well as its reduction due to the square of distance to the source and the angle it hits the detector surface. When the pixel intensities of the original image are corrected by these factors, the image contrast, reduced by the overall attenuation in the exposure phase, are recovered, allowing one to see details otherwise concealed due to the low contrast. In order to verify the soundness of this approach, synthetic images of objects of different shapes, such as plates and tubes, incorporating defects and statistical fluctuation, have been generated, recorded for further comparison and afterwards processed to improve their contrast. The developed algorithm which, generates processes and plots the images has been written in Fortran 90 language. As the resulting final images exhibit the expected improvements, it therefore seemed worthwhile to carry out further tests with actual experimental radiographies.

Enhancing the quality of radiographic images acquired with point-like gamma-ray sources through correction of the beam divergence and attenuation

“…This work originally addressed the experimental verification of the soundness and applicability of a proposed algorithm [7] to unveil concealed feebly attenuating defects enclosed in high attenuating objects inspected by gamma-ray radiography employing a conic beam. Soon it has been realized that the ideal conditions and parameters assumed in the mentioned work -dealing solely with synthetic images -could not be applied, and hence, additional techniques have to be developed to process the real ones.…”

Manufacturing techniques studies of ceramics by neutron and γ-ray radiography

“…In the algorithm used for this purpose [3] the matrix elements of a properly corrected image would have the same value, within the statistical fluctuation and noise, disregarding obviously the heterogeneities due to the rollers and tempers. The approach to perform this correction outlined as follows:…”

“…For sake of simplification they are replaced by apparent average values, which have to be experimentally determined. Further details can be found elsewhere [3,4].…”