Image recovery techniques for x-ray computed tomography in limited data environments

There is an increasing requirement throughout LLNL for nondestructive evaluation using X-ray computed tomography (CT). In many cases, restrictions on data acquisition time, imaging geometry, and budgets make it unfeasible to acquire projection data over enough views to achieve desired spatial resolution using conventional CT methods. In particular, conventional CT methods are non-iterative algorithms that have the advantage of low computational effort, but they are not sufficiently adaptable to incorporate prior information or non-Gaussian statistics. Most currently existing iterative tomography algorithms are based on methods that are time consuming because they converge very flowingly, if at all. The goal of the work was to develop a set of limited data CT reconstruction tools and then demonstrate their usefulness by applying them to a variety of problems of interest to LLNL. In this project they continued their development of reconstruction tools and they have demonstrated their effectiveness on several important problems.