It is a requirement to determine properties such as interfaces and spatial density distributions from radiographic experiments to the highest possible accuracy.
Given the difficulty of the experiments, the radiographic images obtained are subject to a number of factors (such as noise, blur and scatter) that degrade the image quality. This leads to an uncertainty in the solution.
Bayes Law (a probability law) is used to improve the accuracy of solutions by using ‘prior knowledge’ to reject unphysical solutions. AWE has for many years collaborated with Los Alamos National Laboratory in the United States in the development of a software application to do this, called the Bayes Inference Engine. An example of how results can be improved by the inclusion of prior knowledge is shown (Figures 1 & 2).
Figure 1. |
Figure 2. |
It is also important to have a detailed understanding of radiography. This is done though experimental measurement and theoretical prediction, e.g. by using the Monte Carlo software, MCNP, to calculate X-ray scatter.