The analysis and correction of neutron scattering effects in neutron imaging

A method of correcting for the scattering effects present in neutron radiographic and computed tomographic imaging has been developed. Prior work has shown that beam, object, and imaging system geometry factors, such as the LID ratio and angular divergence, Rue the primary sources contributing to the degradation of neutron images. With objects smaller than 20-40 mm (0.75-1.5 in.) in width, a parallel beam approximation can be made where the effects from geometry are negligible. Factors which remain important in the image formation process Rue the pixel size of the imaging system, neutron scattering, the size of the object, the conversion material, and the beam energy spectrum. The Monte Carlo N-Particle transport code, version 4A (MCNP4A), was used to separate and evaluate the effect that each of these parameters has on neutron image data. The simulations were used to develop a correction algorithm which is easy to implement and requires no a priori knowledge of the object. The correction algorithm is based on the determination of the object scatter function (OSF) using available data outside the object to estimate the shape and magnitude of the OSF based on a Gaussian functional form. For objects smaller than 1 mm (0.04 in.) in width, the correction function can be well approximated by a constant function. Errors in the determination and correction of the MCNP simulated neutron scattering component were under five percent and larger errors were only noted in objects which were at the extreme high end of the range of object sizes simulated. The Monte Carlo data also indicated that scattering does not play a significant role in the blurring of neutron radiographic and tomographic images. The effect of neutron scattering on computed tomography is shown to be minimal at best, with the most serious effect resulting when the basic backprojection method is used.