The Effect of Scattered Radiation in X-Ray Techniques—Experiments and Theoretical Considerations
Scattered radiation generated inside a specimen may significantly influence the flaw sensitivity by reducing the relative contrast of the flaw indication . This statement holds if the scattered radiation produces a uniform intensity distribution in the film or detector plane, i.e. it is non-image forming while contributing to the radiographie projection. The introduction of built-up factors yields an appropriate description of the corresponding relative contrast reduction in the radiographie image . In general, the underlying physical process can be treated as an X-ray or photon transport problem [3–6] based on a Boltzmann type equation. This approximation does not need the assumption of a uniform distributed field of scattered radiation. There are several attempts known to solve this problem for NDE applications in terms of Monte Carlo simulation [5, 7–8]. But this technique is only in a qualified sense applicable to practical testing problems with a large variety of factors like 3D object description, finite focal spot, energy dependence of source and interaction mechanisms, and others to be considered requiring a huge number of realizations to receive statistically significant results. Other techniques based on the solution of the corresponding integral transport equation [9–10] employing two stage algorithms. The first stage is known as transport stage where the photon flow resulting from the angular photon sources are computed. The second stage is known as the source computing stage where the scattering sources resulting from the Compton or coherent scattering are computed. Results from these calculations show, that for some cases the scattered radiation does not only decrease the contrast in a radiograph but shows a geometrical dependence which overlays the image from the direct radiation. Finally, an analytical simulation procedure to describe the scattered photon flux was developed which is based on the theory of Markovian processes with random structure [11–12].