Noise and Artifact Reduction in Computed Tomography (CT) Images

Current high-energy computed tomography (CT) techniques for nondestructive testing (NDT) of solid rocket motors (SRM) and other large industrial objects often produce inferior imagery due to artifacts, impairing detection, identification, and measurement of anomalies. Phase I results verified that these artifacts arise from failure of assumptions implicit in the filtered backprojection algorithm. In this Phase II, JDLL extends knowledge of artifact mechanisms and successful corrections developed in Phase I, testing these algorithms against real data, permanently implementing these algorithms on a production high-energy CT system, and producing a commercially viable product that will be useful throughout the industrial high-energy CT NDT community. JDLL’s approach adaptively modifies the individual X-ray measurements in a non-uniform manner, reversing the effect of failure of the underlying CT data acquisition assumptions for each particular application.