الفهرس 
INTRODUCTIONOnly 14 pages are availabe for public view
 |  | from | 86 |  |  |
| | | from | 86 | | |
Abstract
The digital processing of the neutron radiography images gives the possibility for data quantification. In this case an exact relation between the measured neutron attenuation and the real macroscopic attenuation coefficient for every point of the sample is required. The assumption that the attenuation of the neutron beam through the sample is exponential is valid only in an ideal case where a monochromatic beam, non scattering sample and non background contribution are assumed. In the real case these conditions are not fulfilled and in dependence on the sample material we have more or less deviation from the exponential attenuation law. Because of the high scattering cross-sections of hydrogen (as=80.26 barn) for thermal neutrons, the problem with the scattered neutrons at quantitative radiography investigations of hydrogenous materials (as PE, Oil, H20, etc) is not trivial. For these strong scattering materials the neutron beam attenuation is no longer exponential and a dependence of the macroscopic attenuation coefficient on the material thickness and on the distance between the sample and the detector appears. When quantitative radiography (2D) or tomography investigations (3D) are performed, some image correction procedures for a description of the scattering effect are required.
This thesis presents a method that can be used to enhance the neutron radiography image for objects with high scattering materials like hydrogen, carbon and other light materials. This method uses the Monte Carlo code, MCNP5, to simulate the heutron radiography process and get the flux distribution for each pixel of the image and determine the scattered neutrons distribution that causes the image blur and” then subtract it from the initial image to improve its quality.