Fast and Robust Algorithms and Software for Limited-Data Computed Tomography

Nondestructive testing and evaluation (NDT&E) are critical quality control steps when manufacturing munitions and warfare systems. Traditional methods of inspection can identify defects located at or near the surface of an object but cannot reveal internal defects. X-ray and neutron radiography are nonintrusive measurement techniques that record 2D path-integrated images of a target object, called projections. Moreover, the 3D density field needed for comprehensive NDT&E can be reconstructed from a set of projections, each recorded from a unique perspective, using a computed tomography (CT) algorithm. Commercial CT algorithms require hundreds to thousands of projection angles to produce an accurate reconstruction, but the time needed to record this data restricts the use of X-ray or neutron radiography CT for NDT&E to R&D settings. Scientific CT algorithms have been developed for limited-angle CT, which could facilitate fast reconstructions by reducing the requisite number of projections. However, such algorithms are not available in commercial software. To address this need, we will develop the mathematical and software capabilities required for fast and robust limited-angle CT. Our user-friendly software will feature standard and research-grade reconstruction algorithms that have been benchmarked to Army specifications.