Preserving dark skies with neuromorphic camera technology

随着近地轨道(LEO)卫星星座开始形成，卫星条纹污染在广域天文成像中变得越来越普遍。随着数万颗卫星将在未来十年发射，这个问题似乎会变得更加严重。卫星轨迹通常具有很大的不确定性，再加上它们在天空中的快速角运动，因此很难可靠地预测它们是否/何时会进入望远镜的视场。随着星座的存在，天文学家迫切需要预防措施。神经形态(基于事件的)相机技术为这一领域带来了巨大的希望。基于事件的摄像机试图复制生物视网膜的功能，有效地仅记录场景中有意义的信息。它们能够以比基于帧的替代方案低得多的数据速率提供几乎连续的信息流，这对于快速响应系统来说都是有利的特征。我们的目标是开发一种双重用途的仪器，使天文学和空间领域感知(SDA)社区受益，利用神经形态技术来探测卫星并通过提炼现有(通常是过时的)轨道状态信息来补充星表。提供卫星在轨道上污染天文兴趣领域的早期预警。Capability(1)在快速增长的SDA市场具有巨大的商业潜力，英国国家空间战略强调，这是英国政府雄心勃勃地建立全球领导地位的领域。由于狮子座人口的激增，现有的监视网络承受着巨大的压力，人们正在推动开发新的主权SDA传感器。拟议中的项目将探索神经形态摄像机应用于空间监视的能力。具体地说，我们确定了以下主要目标：在我们的校园天文台调试包含最新一代神经形态相机和现有Warwick仪器的原型仪器，随后在我们位于加那利岛拉帕尔马的设施调试。开发数据简化软件、亮度和位置校准、卫星检测和轨迹改进，基于项目负责人在STFC资助的博士生期间开发的现有算法。模拟不同的传感器架构以建立早期预警系统的最佳观测策略，同时考虑卫星污染问题和SDA的要求。概念验证研究，以验证传感器架构模拟的结果与原型仪器，在面临各种情况时，利用编目信息来评估仪器的性能。

Satellite streak contamination is becoming increasingly prevalent in wide-field astronomical imaging, as constellations of satellites are beginning to take shape in low Earth orbit (LEO). The problem looks set to worsen, with tens of thousands of satellites due to launch over the coming decade. The typically large uncertainties in satellite trajectories, coupled with their fast angular motion across the sky, make it challenging to reliably predict if/when they will pass within the field of view of a telescope. With the constellations here to stay, astronomers are urgently in need of preventative measures.Neuromorphic (event-based) camera technology holds great promise for this area. Event-based cameras seek to replicate the functionality of the biological retina, efficiently registering only the meaningful information in a scene. They are capable of providing a near-continuous stream of information with much lower data rates than frame-based alternatives, both favourable characteristics for a rapid-response system.We aim to develop a dual-purpose instrument to benefit both the astronomy and space domain awareness (SDA) communities, utilising neuromorphic technology to:Detect satellites and supplement catalogues by refining existing (often out-of-date) orbital state information.Provide early warning of satellites on course to contaminate an astronomical field of interest.Capability (1) has significant commercial potential within the rapidly growing SDA market, highlighted in the National Space Strategy as an area in which the UK government has ambitions to establish global leadership. Existing surveillance networks are under immense strain due to the soaring population in LEO and there is a drive to develop new sovereign SDA sensors.The proposed project will explore the capabilities of neuromorphic cameras when applied to space surveillance. Specifically, we identify the following key objectives:Commissioning of a prototype instrument comprising a latest-generation neuromorphic camera and existing Warwick instrumentation at our campus observatory, and subsequently at our facilities in La Palma, Canary Islands.Development of software for data reduction, calibrations for brightness and position, satellite detection, and trajectory refinement, building from existing algorithms developed during the Project Lead's STFC-funded PhD studentship.Simulation of different sensor architectures to establish an optimal observational strategy for the early warning system, taking both the satellite contamination problem and requirements for SDA into consideration.Proof-of-concept study to validate the findings from the sensor architecture simulation with the prototype instrument, making use of catalogued information to assess the instrument's performance when faced with a variety of scenarios.