Efficient Onboard Multitask AI Architecture Based on Self-Supervised Learning

There is growing interest toward the use of artificial intelligence (AI) directly onboard satellites for quick analysis and rapid response to critical events such as natural disasters. This article presents a blueprint to the mission designer for the development of a modular and efficient deep learn...

Full description

Saved in:
Bibliographic Details
Main Authors: Gabriele Inzerillo, Diego Valsesia, Enrico Magli
Format: Article
Language:English
Published: IEEE 2025-01-01
Series:IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing
Subjects:
Online Access:https://ieeexplore.ieee.org/document/10758783/
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1850258307387752448
author Gabriele Inzerillo
Diego Valsesia
Enrico Magli
author_facet Gabriele Inzerillo
Diego Valsesia
Enrico Magli
author_sort Gabriele Inzerillo
collection DOAJ
description There is growing interest toward the use of artificial intelligence (AI) directly onboard satellites for quick analysis and rapid response to critical events such as natural disasters. This article presents a blueprint to the mission designer for the development of a modular and efficient deep learning payload to address multiple onboard inference tasks. In particular, we design a self-supervised lightweight backbone that provides features to efficient task-specific heads. The latter can be developed independently and with reduced data labeling requirements thanks to the frozen backbone. Experiments on three sample tasks of cloud segmentation, flood detection, and marine debris classification on a 7-W embedded system show competitive results with inference quality close to high-complexity state-of-the-art models and high throughput in excess of 8 Megapixel/s.
format Article
id doaj-art-5197dc4d3d2341dd830a2febcd4bdc3e
institution OA Journals
issn 1939-1404
2151-1535
language English
publishDate 2025-01-01
publisher IEEE
record_format Article
series IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing
spelling doaj-art-5197dc4d3d2341dd830a2febcd4bdc3e2025-08-20T01:56:13ZengIEEEIEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing1939-14042151-15352025-01-011882883810.1109/JSTARS.2024.350277610758783Efficient Onboard Multitask AI Architecture Based on Self-Supervised LearningGabriele Inzerillo0https://orcid.org/0009-0009-2027-0562Diego Valsesia1https://orcid.org/0000-0003-1997-2910Enrico Magli2https://orcid.org/0000-0002-0901-0251Department of Electronics and Telecommunications, Politecnico di Torino, Torino, ItalyDepartment of Electronics and Telecommunications, Politecnico di Torino, Torino, ItalyDepartment of Electronics and Telecommunications, Politecnico di Torino, Torino, ItalyThere is growing interest toward the use of artificial intelligence (AI) directly onboard satellites for quick analysis and rapid response to critical events such as natural disasters. This article presents a blueprint to the mission designer for the development of a modular and efficient deep learning payload to address multiple onboard inference tasks. In particular, we design a self-supervised lightweight backbone that provides features to efficient task-specific heads. The latter can be developed independently and with reduced data labeling requirements thanks to the frozen backbone. Experiments on three sample tasks of cloud segmentation, flood detection, and marine debris classification on a 7-W embedded system show competitive results with inference quality close to high-complexity state-of-the-art models and high throughput in excess of 8 Megapixel/s.https://ieeexplore.ieee.org/document/10758783/Multitask learningonboard AIself-supervised learning (SSL)
spellingShingle Gabriele Inzerillo
Diego Valsesia
Enrico Magli
Efficient Onboard Multitask AI Architecture Based on Self-Supervised Learning
IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing
Multitask learning
onboard AI
self-supervised learning (SSL)
title Efficient Onboard Multitask AI Architecture Based on Self-Supervised Learning
title_full Efficient Onboard Multitask AI Architecture Based on Self-Supervised Learning
title_fullStr Efficient Onboard Multitask AI Architecture Based on Self-Supervised Learning
title_full_unstemmed Efficient Onboard Multitask AI Architecture Based on Self-Supervised Learning
title_short Efficient Onboard Multitask AI Architecture Based on Self-Supervised Learning
title_sort efficient onboard multitask ai architecture based on self supervised learning
topic Multitask learning
onboard AI
self-supervised learning (SSL)
url https://ieeexplore.ieee.org/document/10758783/
work_keys_str_mv AT gabrieleinzerillo efficientonboardmultitaskaiarchitecturebasedonselfsupervisedlearning
AT diegovalsesia efficientonboardmultitaskaiarchitecturebasedonselfsupervisedlearning
AT enricomagli efficientonboardmultitaskaiarchitecturebasedonselfsupervisedlearning