Year
2024
Abstract
| Warhead confirmation could be a component of future arms control treaties, and may include confirmation measurements based on geometric attributes. Mature passive neutron imaging systems, such as the Neutron Coded Aperture (NCA) jointly developed by Oak Ridge National Laboratory and Sandia National Laboratories, are able to test for the presence of highly specific attributes based on the geometric information of both fast and thermal neutron sources within an assembly. Here, we present an analysis for the example attribute: “an assembly containing a physically extended fast-neutron source, completely enclosed by a moderating material.” Our analysis acts on the set of pixels in an image reconstructed from simulated data. We use a combination of smoothing, thresholding, and clustering to identify individual fast and thermal neutron sources within the reconstructed image. The extent of a fast-neutron source is tested using the weighted covariance of the fast-neutron source pixels. Enclosure by moderating material is tested for by requiring that each pixel in a fast neutron source is overlapped by a thermal-neutron source. In addition to the positive case, we challenge our attribute analysis using several image datasets that are generated from simulations of randomized assemblies that do not possess the attribute. We find that we are able to confirm positive images with a high probability, while rejecting negative images with a false positive rate of less than 25% for each of the challenge datasets investigated. |