Vicente Westerhout, Pontificia Universidad Católica de Valparaíso; Sebastián Valdivia, Pontificia Universidad Catolica de Valparaiso; Esteban Vera, Pontificia Universidad Catolica de Valparaiso
Keywords: Space domain awareness, event-based camera.
Abstract:
In the last couple of years, satellites residing in Low-Earth Orbits (LEO) have greatly increased in numbers and are expected to exponentially grow in the near future. Therefore, it is imperative for Space Situational Awareness (SSA) systems to be up to the task of detecting and maintaining ephemeris of the ever growing number of objects, such as satellites or debris, to help preserve Earths near space safety, avoiding collision hazards. This duty has been transferred, slowly, from radar sites to optical telescopes to reduce costs while increasing the coverage. Nowadays there are many efforts to design efficient optical-based SSA systems with components-of-the-shelf (COTS), benefiting the community by having more eyes in the sky to detect and track objects of interest.
However, observing objects in LEO with optical telescopes also has many restraints. For instance, it can only be done for a few hours every night, weather allowing. Moreover, there are technical showstoppers too. LEO satellite monitoring requires capturing the largest portion of the sky as possible and the longest time without interruptions, considering the amount of objects and their different orbits (inclinations) and speeds. Also, COTS optical telescopes are mostly made for astronomical observations so they often have long focal lengths, meaning small Field of Views (FOV). When coupled to modern CMOS cameras with large pixel counts, the continuous monitoring of the sky at shorter exposures in contrast with astronomical applications, leads to saturation in terms of transmission, processing and storage of data. Therefore, given the size and brightness of the space objects of interest, optical SSA solutions have to be designed with the right balance between FoV, pixel size, pixel count, aperture size, and exposure time. Notice, though, that we are also photographing the stars, which are not necessarily informative for SSA purposes.
On the other hand, event based sensors (EBS) have emerged as a solution to intensity detectors, being more efficient to detect sparse changes in the scenes at higher frame rates and at worse SNR conditions. The asynchronous operation of its pixels avoids the redundant elements of the scene, that is, they record only moving objects in the scene. Also the temporal resolution of 1 microsecond makes them a perfect device for surveillance tasks. Furthermore, the low data rates provided by capturing only the changes in the sky reduce the need for transmission, storage and computer processing.
Even though event based SSA solutions reduce the big data workload associated with traditional imaging detectors, we still have to understand how to properly use them for this task. State-of-the-art works have shown object detections during night and day including stars, planets, satellites and debris. Although, most detections are made using telescopes with large focal lengths, i.e., small FOV which do not solve the needs of SSA. On the other hand, most approaches had focused on the software part of the systems, leaving behind the opportunity to improve the acquisition by tuning the correct parameters and enhancing the implementation setup. At last, another approach assesses the sensitivity of EBSs in capturing night sky objects orienting the workflow towards the comparison between two models of these sensors. However, as far as we know none of the state-of-the-art work analyzes the correct way in which an event-based SSA system must be implemented, meaning, little research has been made to define which are the proper optics to capture determined objects of interest.
Therefore, through simulations and experiments, we explore the use of EBC for the detection of LEO objects by using different implementations of the optical system (aperture, focal-length, FoV, etc.) with the purpose of better understanding the trade off between the selection of a specific FOV and the detection ability of the sensor. This task is necessary to properly design an all-sky event based SSA system, most likely using an array of small telescopes. Through simultaneous experiments using different wide and narrow FoV telescopes and photographic lenses, we are currently measuring characteristics (such as apparent magnitude and orbital regime) that can determine the detection limits for each implementation at the same observing conditions, which can be verified later when comparing with available objects catalogs. Then, a multiscale analysis of these characteristics are being made to optimize the design of an event-based SSA application; where the main benefits are that the detection of objects of interest can occur with a low power and low workload in contrast with traditional image processing approaches.
Date of Conference: September 19-22, 2023
Track: SDA Systems & Instrumentation