Improving the quality of radiographic images acquired with conical radiation beams through divergence correction and filtering

Earlier works have shown the feasibility to correct the deformation of the attenuation map in radiographs acquired with conical radiation beams provided that the inspected object could be expressed into analytical geometry terms. This correction reduces the contribution of the main object in the rad...

Full description

Saved in:
Bibliographic Details
Published inJournal of physics. Conference series Vol. 630; no. 1; pp. 12013 - 12021
Main Authors Silvani, M I, Almeida, G L, Latini, R M, Bellido, A V B, Souza, E S, Lopes, R T
Format Journal Article
LanguageEnglish
Published Bristol IOP Publishing 15.07.2015
Subjects
Online AccessGet full text

Cover

Loading…
More Information
Summary:Earlier works have shown the feasibility to correct the deformation of the attenuation map in radiographs acquired with conical radiation beams provided that the inspected object could be expressed into analytical geometry terms. This correction reduces the contribution of the main object in the radiograph, allowing thus the visualization of its otherwise concealed heterogeneities. However, the non-punctual character of the source demanded a cumbersome trial-and-error approach in order to determine the proper correction parameters for the algorithm. Within this frame, this work addresses the improvement of radiographs of specially tailored test-objects acquired with a conical beam through correction of its divergence by using the information contained in the image itself. The corrected images have afterwards undergone a filtration in the frequency domain aiming at the reduction of statistical fluctuation and noise by using a 2D Fourier transform. All radiographs have been acquired using 165Dy and 198Au gamma-ray sources produced at the Argonauta research reactor in Institutode Engenharia Nuclear - CNEN, and an X-ray sensitive imaging plate as detector. The processed images exhibit features otherwise invisible in the original ones. Their processing by conventional histogram equalization carried out for comparison purposes did not succeed to detect those features.
ISSN:1742-6588
1742-6596
DOI:10.1088/1742-6596/630/1/012013