An Edge Computing Algorithm for Onboard Processing of Electro-Optical Imagery

Matthew Britton, The Aerospace Corporation; Christopher Griffith, The Aerospace Corporation; Brian Baptista, The Aerospace Corporation

Keywords: on-board processing, electro-optic, image processing, autonomy, sensors

Abstract:

This paper presents techniques enabled by a real-time embedded algorithm for onboard processing of electro-optic rate-track imagery acquired by ground and space sensors.  These techniques offer many advantages: 104 reduction in data volume to be transferred off-sensor, elimination of confusion between RSOs and stars, multi-target tracking, multi-frame integration of both stars and RSOs at the single pixel-level, a variety of electro-optic feedback control options, and absolute metric and photometric calibration.  These advantages are available in real time at rates exceeding video frame rates.  These techniques promote sensor autonomy and alleviate the restrictions of communication bandwidth and processing latency that arise when transferring raw imagery off-sensor.  This paper demonstrates these techniques using data collected on a GEO target from a 30cm ground telescope and through an ongoing embedded development effort using a Zybo Z7-20 FPGA development board. These techniques are patent pending and available for licensing from The Aerospace Corporation for government and commercial use.

Date of Conference: September 19-22, 2023

Track: SDA Systems & Instrumentation

View Paper