NDTnet 1998 Aug, Vol.3 No.8

Abstract

Practical applications of tomographic methods for NDT of wide range of industrial products require different approaches to input data collecting and image producing. In many cases it is enough to get plane images to detect possible defects in some products. It is known and widely used a method, when a set of plane images provides forming of 3D model of inspected object. The method is informative enough, but in some cases it requires too much time to collect input data.

The known formulas of Radon transform for odd-D spaces allow to implement fundamentally different approaches. Methods based on these formulas provide the reconstruction of 3D images using projections collected by plane set of 2D type detectors under 3D configuration of penetrate emission of the source. In this connection there are different possible scanning schemes, geometric shapes of penetrative beam, constructions of 2D plane of detectors. Beside this, there are a number of factors, which impact on obtaining results, their quality and especially space and density resolution.

For collecting data for X-ray tomography the best characteristics are shown by conebeam tomographs. They have some important advantages: 1) efficient use the solid angle of X-ray emission; 2) significant reduction in the time needed to collect a sufficient number of data; 3) elimination of the inaccuracy due to misalignment of cross sectional images; 4) the divergent shape of the beam provides facilities for magnification in the distances of the source to detector and of the sources to the axis of rotation, which used in conjunction with a microfocus X-ray source opens the way to high resolution.

To analyze the variety of approaches to 3D reconstruction we select two main groups.
1) Methods used complete scanning geometry (which meet the Kirillov-Tuy or another completeness conditions). They based on some exact closed inversion formulas. The best characteristics in this group are shown by methods of Smith and Grangeat. Their structure has the form of convolution backprojection and they differ by the form and parameters of the inversion formulas. In spite of sufficient computational volume they gives good image quality, which is restricted only by technical limitations.
2) Methods used incomplete scanning geometry (circle). Usually some approximations were made in their derivation. This fact results in necessity to solve extrapolation problem to reduce artifacts due incomplete data set. The most efficient and popular method from this group is algorithm of Feldkamp. It has high computational efficiency and gives acceptable results when using cone-beams of a small divergence angle.

Abstract Source:
Book of Abstracts, 7th European Conference on Non-Destructive Testing, 26-29 May 1998, ISBN: 87-986898-0-00

Full-Text Source:
Proceedings of the 7th European Conference on Non-Destructive Testing, 26-29 May 1998, ISBN:

Publication Contact:
7th ECNDT, Park Alle 345, DK-2605 Broendby, Denmark, Fax: +45 46 26 70 11, Email: 101373.3414@compuserve.com