SelfieStick: Towards Earth Imaging from a Low-Cost Ground Module Using LEO Satellites
SelfieStick: Towards Earth Imaging from a Low-Cost Ground Module Using LEO Satellites
复制标题
DOI：
10.1109/ipsn54338.2022.00025
复制
发表时间：
2022-05
期刊：
影响因子：
--
通讯作者：
Vaibhav Singh;Osman Yağan;Swarun Kumar
中科院分区：
文献类型：
--
作者：
Vaibhav Singh;Osman Yağan;Swarun Kumar
从手持设备实时获取头顶低地球轨道(LEO)卫星图像可以产生革命性的应用：跟踪野火、自然灾害和天气事件。今天，从头顶上的LEO卫星实时获取图像是具有挑战性的。LEO卫星地面接收器体积庞大、价格昂贵，在世界上部署稀少。尽管如今围绕地球轨道运行的低轨小卫星呈指数级增长--但在这种卫星上拍摄的图像与偏远和生态敏感地区最需要图像的用户之间存在着巨大的时间差距。本文介绍了一种新的卫星接收器系统--Sselfiestick，它探索使用一个低成本(30英镑)的微型接收器来减少获取实时卫星图像数据的障碍。Sigfiestick的核心方法利用了架空低地球轨道卫星的多样性，因为它们近年来呈指数级增长。虽然来自这类卫星的信号可能单独很弱，特别是在低成本的接收器上，但Selfiestick将包含许多这样的卫星上地球同一部分的潜在图像的嘈杂的射频捕获拼接在一起，以生成干净的地球图像。这是通过在考虑到不同卫星视角、其轨道和无线信道的情况下对RF信号进行适当的变换和对准之后，在RF域(而不是传统的图像域)中组合弱信号来实现的。在RTL-SDR平台上对来自NOAA星座的卫星捕获进行了详细的实验评估，结果表明，通过合并10颗卫星的图像，PSNR提高了5dB。
Real-time access to overhead Low-Earth Orbit (LEO) satellite imagery from a handheld device can have transformative applications: tracking wild-fire, natural disasters and weather events. Today, real-time access to images from LEO satellites overhead is challenging to obtain. LEO satellite ground receivers are bulky, expensive and sparsely deployed in the world. Despite the exponential increase in LEO small satellites orbiting the planet today-there is a significant time gap between an image capture on such a satellite and users who need it the most in remote and ecologically-sensitive regions. This paper presents SelfieStick, a novel satellite receiver system that explores reducing this barrier of access to real-time satellite imagery data using a single low cost (< $ 30) tiny receiver. SelfieStick's core approach takes advantage of the multiplicity of overhead Low-Earth Orbit satellites due to their exponential rise in recent years. While signals from such satellites may be individually weak, especially at a low-cost receiver, SelfieStick stitches together noisy RF captures containing underlying images of the same part of the Earth across many such satellites to generate clean Earth images. This is made possible by combining weak signals in the RF domain (rather than the traditional image domain) after appropriately transforming and aligning the RF signals accounting for different satellite perspectives, their orbits and wireless channels. A detailed experimental evaluation on the RTL-SDR platform on satellite captures from the NOAA constellation demonstrates a PSNR improvement of 5 dB through combining of images across 10 satellites.