AI/ML-Enhanced Quantitative X-ray for Rapid Identification of Projectile Fill and Assembly Defects

During the proposed Phase I effort, ZKxKZ will design, develop and demonstrate a new hardware/software system to automate inspection and interpretation of X-ray images of energetic projectiles as a routine part of the manufacturing process. Phase I work will focus on a system capable of evaluating projectiles up to 50mm, but the technology will ultimately be scalable to 155mm projectiles. We will place particular emphasis on evaluation of the Excalibur round. The proposed system will feature Artificial Intelligence and Machine Learning (AI/ML) to automate the interpretation of X-ray images. We will use our existing FRAG-X AI/ML-driven quantitative X-ray inspection hardware and software (developed with Army and Air Force SBIR funding to automate the location and characterization of fragments after a warhead arena test) as a starting point. A key feature of Phase I will be the development of a full safety plan that addresses both handling of energetic materials, and X-ray system safety. Our work will be significantly advanced by participation of teammates from Leidos (who has already fielded an 155mm X-ray inspection system) and Raytheon (manufacturer of the Excalibur round).