Development of a minaturised forwards looking imager using deep learning for responsive operations

Steve Greenland, Murray Ireland, Chisato Kobayashi, Peter Mendham, Mark Post, David White

Research output: Chapter in Book/Report/Conference proceedingConference contribution book

Abstract

This work presents the design and prototyping work of a miniaturised camera system with integrated ‘deep learning’ neural network capabilities, developed within a framework for implementing autonomous data processing onboard small and nanosatellites. The framework targets low-resource algorithms developed in other sectors including autonomous vehicles and commercial machine learning. For proof of concept, the system has been initially trained for real time cloud detection and classification, looking 1 min ahead of the satellite to enable responsive decision making for Earth observation and telecommunication applications. The design has been miniaturised and modularised to allow accommodation on small and nanosatellite systems. Flight representative and heritage components have been selected for prototyping. Compatibility of the autonomy framework with ECSS and CCSDS standards and existing off-the-shelf flight software was evaluated. A simulator to facilitate end to end testing of the system has been developed using existing data sets as input, incorporating distortions to test robustness. Results show that a competitive low power < 2 W system can be delivered, with the chain < 5 seconds from capture to input into the onboard planning and with timing consistent with continuous real time decision-making.
LanguageEnglish
Title of host publication4S Symposium 2018
Subtitle of host publicationThe Symposium on Small Satellites for Earth Observation
Place of PublicationNoordwijk
Number of pages9
Publication statusPublished - 28 May 2018

Fingerprint

Nanosatellites
Image sensors
Decision making
Telecommunication
Learning systems
Simulators
Earth (planet)
Cameras
Satellites
Neural networks
Planning
Testing
Deep learning

Keywords

  • onboard operations
  • miniaturised camera systems
  • deep learning
  • neural networks
  • nanosatellites
  • autonomous vehicles

Cite this

Greenland, S., Ireland, M., Kobayashi, C., Mendham, P., Post, M., & White, D. (2018). Development of a minaturised forwards looking imager using deep learning for responsive operations. In 4S Symposium 2018: The Symposium on Small Satellites for Earth Observation Noordwijk.
Greenland, Steve ; Ireland, Murray ; Kobayashi, Chisato ; Mendham, Peter ; Post, Mark ; White, David. / Development of a minaturised forwards looking imager using deep learning for responsive operations. 4S Symposium 2018: The Symposium on Small Satellites for Earth Observation. Noordwijk, 2018.
@inproceedings{ba03ab5aa67844368110271c93cdf911,
title = "Development of a minaturised forwards looking imager using deep learning for responsive operations",
abstract = "This work presents the design and prototyping work of a miniaturised camera system with integrated ‘deep learning’ neural network capabilities, developed within a framework for implementing autonomous data processing onboard small and nanosatellites. The framework targets low-resource algorithms developed in other sectors including autonomous vehicles and commercial machine learning. For proof of concept, the system has been initially trained for real time cloud detection and classification, looking 1 min ahead of the satellite to enable responsive decision making for Earth observation and telecommunication applications. The design has been miniaturised and modularised to allow accommodation on small and nanosatellite systems. Flight representative and heritage components have been selected for prototyping. Compatibility of the autonomy framework with ECSS and CCSDS standards and existing off-the-shelf flight software was evaluated. A simulator to facilitate end to end testing of the system has been developed using existing data sets as input, incorporating distortions to test robustness. Results show that a competitive low power < 2 W system can be delivered, with the chain < 5 seconds from capture to input into the onboard planning and with timing consistent with continuous real time decision-making.",
keywords = "onboard operations, miniaturised camera systems, deep learning, neural networks, nanosatellites, autonomous vehicles",
author = "Steve Greenland and Murray Ireland and Chisato Kobayashi and Peter Mendham and Mark Post and David White",
note = "Title changed from {"}Design & prototyping of a minaturised forwards looking imager using deep learning for responsive onboard operations{"} to {"}Development of a miniaturised forwards looking imager using deep learning for responsive onboard operations{"} after acceptance.",
year = "2018",
month = "5",
day = "28",
language = "English",
booktitle = "4S Symposium 2018",

}

Greenland, S, Ireland, M, Kobayashi, C, Mendham, P, Post, M & White, D 2018, Development of a minaturised forwards looking imager using deep learning for responsive operations. in 4S Symposium 2018: The Symposium on Small Satellites for Earth Observation. Noordwijk.

Development of a minaturised forwards looking imager using deep learning for responsive operations. / Greenland, Steve; Ireland, Murray; Kobayashi, Chisato; Mendham, Peter; Post, Mark; White, David.

4S Symposium 2018: The Symposium on Small Satellites for Earth Observation. Noordwijk, 2018.

Research output: Chapter in Book/Report/Conference proceedingConference contribution book

TY - GEN

T1 - Development of a minaturised forwards looking imager using deep learning for responsive operations

AU - Greenland, Steve

AU - Ireland, Murray

AU - Kobayashi, Chisato

AU - Mendham, Peter

AU - Post, Mark

AU - White, David

N1 - Title changed from "Design & prototyping of a minaturised forwards looking imager using deep learning for responsive onboard operations" to "Development of a miniaturised forwards looking imager using deep learning for responsive onboard operations" after acceptance.

PY - 2018/5/28

Y1 - 2018/5/28

N2 - This work presents the design and prototyping work of a miniaturised camera system with integrated ‘deep learning’ neural network capabilities, developed within a framework for implementing autonomous data processing onboard small and nanosatellites. The framework targets low-resource algorithms developed in other sectors including autonomous vehicles and commercial machine learning. For proof of concept, the system has been initially trained for real time cloud detection and classification, looking 1 min ahead of the satellite to enable responsive decision making for Earth observation and telecommunication applications. The design has been miniaturised and modularised to allow accommodation on small and nanosatellite systems. Flight representative and heritage components have been selected for prototyping. Compatibility of the autonomy framework with ECSS and CCSDS standards and existing off-the-shelf flight software was evaluated. A simulator to facilitate end to end testing of the system has been developed using existing data sets as input, incorporating distortions to test robustness. Results show that a competitive low power < 2 W system can be delivered, with the chain < 5 seconds from capture to input into the onboard planning and with timing consistent with continuous real time decision-making.

AB - This work presents the design and prototyping work of a miniaturised camera system with integrated ‘deep learning’ neural network capabilities, developed within a framework for implementing autonomous data processing onboard small and nanosatellites. The framework targets low-resource algorithms developed in other sectors including autonomous vehicles and commercial machine learning. For proof of concept, the system has been initially trained for real time cloud detection and classification, looking 1 min ahead of the satellite to enable responsive decision making for Earth observation and telecommunication applications. The design has been miniaturised and modularised to allow accommodation on small and nanosatellite systems. Flight representative and heritage components have been selected for prototyping. Compatibility of the autonomy framework with ECSS and CCSDS standards and existing off-the-shelf flight software was evaluated. A simulator to facilitate end to end testing of the system has been developed using existing data sets as input, incorporating distortions to test robustness. Results show that a competitive low power < 2 W system can be delivered, with the chain < 5 seconds from capture to input into the onboard planning and with timing consistent with continuous real time decision-making.

KW - onboard operations

KW - miniaturised camera systems

KW - deep learning

KW - neural networks

KW - nanosatellites

KW - autonomous vehicles

M3 - Conference contribution book

BT - 4S Symposium 2018

CY - Noordwijk

ER -

Greenland S, Ireland M, Kobayashi C, Mendham P, Post M, White D. Development of a minaturised forwards looking imager using deep learning for responsive operations. In 4S Symposium 2018: The Symposium on Small Satellites for Earth Observation. Noordwijk. 2018