Approaches for astrometry using event-based sensors

Event-based sensors are novel optical imaging devices that offer a different paradigm in which to image space and resident space objects. Also known as silicon retinas, these custom silicon devices make use of independent and asynchronous pixels which produce data in the form of events generated in response to changes in log-illumination rather than in the conventional frames produced by CCD-based imaging sensors. This removes the need for fixed exposure times and frame rates but requires new approaches to processing and interpreting the spatio-temporal data produced by these sensors. The individual nature of each pixel also yields a very high dynamic range, and the asynchronous operation provides a high temporal resolution. These characteristics make event-based cameras well suited to terrestrial and orbital space situational awareness tasks. Our previous work with these sensors highlighted the applicability of these devices for detecting and tracking resident space objects from LEO to GEO orbital regimes, both during the night and daytime without modification to the camera or optics. Building upon this previous work in applying these artificial vision systems to space situational awareness tasks, we present a study into approaches for calculating astrometry from the event-based data generated with these devices. The continuous nature of these devices, and their ability to image whilst moving, allows for new and computationally efficient approaches to astrometry, applicable both to high-speed tracking from terrestrial sensors and low-power imaging from orbital platforms. Using data collected during multiple sets of telescope trials involving co-collections between a conventional sensor and multiple event-based sensors, a system capable of identifying stars and positional information whilst simultaneously tracking an object is presented. Two new prototype event-based sensors, offering increased spatial resolution and higher sensitivity, were also used and characterized in the trial, and updated observation results from these improved sensors are presented. These results further demonstrate and validate the applicability and opportunities offered by event-based sensors for space situational awareness and orbital applications.