The Space Development Agencyâs (SDA) Tracking Layer is tasked with providing global indications, warning, tracking, and targeting of advanced missile threats, including hypersonic missile systems. They are currently leading the development of an Overhead Persistent Infrared (OPIR) Missile Tracking satellite constellation in LEO as part of the DoDâs future threat-driven National Defense Space Architecture (NDSA). Satellite imagery from sensors in LEO typically contain complex cluttered backgrounds accentuated by non-stationary platform motion making it even more challenging to discern new threats of interest. On this effort, Toyon Research Corporation and the Air Force Institute of Technology (AFIT) propose to leverage the AFIT Scene and Sensor Emulation Tool (ASSET) for modeling sensors in LEO and simulating relevant data for algorithm development and testing. Algorithm development will entail methods using novel vision-based geometry for scene registration and parallax mitigation necessary for clutter suppression in data from non-stationary sensors. We will also develop algorithms for automated recognition and acquisition of new threat scenarios leveraging track-before-detect and deep learning architectures. Lastly, algorithms will be optimized to enable integration with future operational systems including on-orbit processing hardwa