Lightcurve generation using neuromorphic event-based sensors

Frame-based optical sensors have been used for many years to produce lightcurves of space objects, from which information about those objects may be inferred. A relatively new type of optical sensor, known as a neuromorphic, or event-based, sensor is gaining attention as a promising tool for space object detection and characterisation. Comprising an array of pixels that function entirely independently and asynchronously, event-based sensors operate with microsecond temporal resolution, high dynamic range, and often with far lower data rates than frame-based sensors. An event-based pixel only outputs data when there is a change of brightness incident on that pixel that exceeds adjustable thresholds. We used a Prophesee Gen 4 EM event-based sensor that outputs not only the time of each detected change in brightness, but also a time-encoded value proportional to the brightness at the pixel after the change, to produce lightcurves of five satellites. The temporal resolution of the lightcurves that we produced varied from object to object, but in all cases there were many data points per second. Furthermore, since the event-based sensor operates without a shutter or set exposure interval, sudden or short duration brightness changes that would otherwise be missed during frame readout or smoothed by the effect of an exposure interval are captured. The results demonstrate that event-based sensors have considerable potential as space situational awareness tools.