Processing of Space Object Data from Optical Observers for Space Domain Awareness

Abstract

This paper presents a novel approach to identify, localize, and associate resident space objects (i.e. Earth-orbiting satellites) across a series of unresolved optical images, to serve as an input to the process of initial and/or precise orbit determination. The images will contain stars and (potentially) satellites, which generally will blend together. The approach involves a sequence of steps including image processing, centroiding, registration, star elimination, and object association. The goal is to develop algorithms that are compact and lend themselves to onboard operation, so that local space domain awareness can be performed by any spacecraft with a taskable camera. Results are presented using imagery obtained from a ground sensor, with satellites in various types of orbits being imaged.