Pith. sign in

REVIEW 3 major objections 6 minor 1 cited by

Advancing Earth Observation: A Survey on AI-Powered Image Processing in Satellites

T0 review · 3 major / 6 minor · reviewed 2026-08-10 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗ Review JSON

Pith's one-line read This survey claims that running pre-trained machine-learning models on board Earth observation satellites is feasible and that the decisive constraints are power, processing capability, memory, and radiation.

desk verdict A useful but non-systematic survey of on-board AI for EO satellites, with an unverified claim to be the first dedicated review. read the letter →

arxiv 2501.12030 v1 pith:TOQJQNVQ submitted 2025-01-21 cs.LG cs.CV

Aidan Duggan , Bruno Andrade , Haithem Afli This is my paper · ORCID
classification cs.LGcs.CV
keywords earthobservationon-boardimageprocessingedgecomputingdeeplearningmodeloptimisationquantizationradiationhardeningsmallsatellites
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

This survey argues that the bottleneck for Earth observation satellites is no longer image capture but getting the data to the ground, and that running pre-trained machine-learning models on board is the workable fix. It identifies four constraints — power, processing capability, memory, and radiation — as the most significant barriers, and reviews mitigation strategies for each. The paper claims to be the first dedicated, up-to-date review of on-board AI image processing for Earth observation, distinct from broader reviews of machine learning in space or satellite operations. If the synthesis is right, engineers can use it as a practical map for choosing hardware, model architectures, and compression techniques for orbital edge computing.

What carries the argument

The organising mechanism is a constraint-to-mitigation mapping. Each of the four constraints — power, processing capability, memory, radiation — is paired with concrete mitigation families: power budgeting and low-power accelerators; reconfigurable and hybrid computing (FPGA plus SoC); lightweight model architectures and optimisation/compression techniques; and radiation-hardened component selection or tolerance testing. The review also uses the CloudScout deployment on PhiSat-1 as the reference case that shows the mapping working in orbit.

What would settle it

A systematic search of the literature with explicit databases, inclusion criteria, and a date cutoff that surfaces a substantial body of dedicated on-board EO image-processing work published before January 2025 would falsify the novelty claim; a documented satellite mission in which a constraint other than power, processing, memory, or radiation proved dominant would falsify the constraint ranking.

Watch

Extended reading notes

Core claim

The central claim is that deploying pre-trained ML models to Earth observation satellites for image processing is feasible today, provided the design accounts for four dominant constraints: limited power, limited processing capability, limited memory, and radiation exposure. For each constraint the paper collects the current mitigation: low-power AI accelerators and system-on-chips for compute, lightweight CNN architectures and optimisation techniques (quantization, pruning, knowledge distillation, layer decomposition) for power and memory, on-board cloud filtering and pre-processing to reduce stored and transmitted data, and radiation-hardened or radiation-tolerant component choices. The mission proof point is CloudScout, a cloud-detection CNN that ran on a low-power vision processing unit aboard the PhiSat-1 mission, performing an inference in 325 ms at 1.8 W with a 2.1 MB model footprint. The paper concludes that real-world on-board deployments remain rare and that quantifying the power and memory savings is the main open research need.

Load-bearing premise

The load-bearing premise is that the papers the survey chose to discuss are representative and complete enough to support both the ranking of the four constraints and the claim that no earlier dedicated review exists.

Editorial extensions

If this is right

  • If the constraint map is correct, on-board filtering of cloudy or low-value images can cut downlink volume and transmission power, making small-satellite missions more responsive.
  • Lightweight architectures and optimisation techniques such as quantization can shrink model footprint and inference power enough for CubeSat-class hardware, with reported power reductions up to 87% in profiling studies.
  • Hybrid and reconfigurable processors (FPGA plus SoC) offer a path to combine radiation tolerance with the compute needed for neural-network inference, at the cost of longer development time than commercial off-the-shelf parts.
  • The small number of in-orbit deployments means the field's next step is measurement: quantifying how much power and memory on-board processing actually saves per mission.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

  • Editorial inference: the survey's own evidence suggests the power saved by not transmitting full images can offset or exceed the power spent on inference, but the paper does not quantify this trade; a direct energy-accounting study on a representative small satellite would settle it.
  • Editorial inference: because the review locates few real deployments, its ranking of the four most significant constraints is provisional; a future mission that names thermal management or data quality as the binding limit would revise the map.
  • Editorial inference: the partial-processing idea noted as unexplored in the paper — transmitting intermediate neural-network layer outputs instead of full images — could be tested today on edge hardware and would give a quantitative handle on memory and bandwidth savings.
Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Referee Report

3 major / 6 minor

Summary. This manuscript is a survey of AI-based image processing on-board Earth observation (EO) satellites. It identifies four constraints (power, processing capability, memory, and radiation) and reviews strategies to mitigate them, including low-power hardware (AI accelerators, SoCs), efficient CNN architectures, model optimization techniques (pruning, quantization, knowledge distillation, layer decomposition), and on-board cloud detection/filtering. The paper claims to be the first dedicated and up-to-date review of deploying pre-trained ML models on-board EO satellites for image processing. It does not present new experiments, measurements, or derivations; its contribution is exclusively a synthesis of prior work.

Significance. If the survey's coverage is trustworthy, it provides a useful entry point for practitioners: it organizes a scattered literature into a constraints/mitigations structure, highlights exemplary deployments such as CloudScout on the PhiSat-1 mission, and points to concrete hardware and model options. The descriptions of individual works appear consistent with the cited sources, and the paper gives credit to a reasonable sample of relevant research. However, the absence of a documented search methodology makes the central claims of being 'thorough' and 'up-to-date' unverifiable, and the novelty claim (being the first dedicated review) is not justified by any systematic evidence. The paper also relies on the authors' own prior result for a key quantitative claim about quantization efficiency. These issues directly affect the survey's reliability and its value to readers.

major comments (3)
  1. [Section 1 and Abstract] The survey claims to be 'up-to-date and thorough' (Abstract) and states that 'to the best of our knowledge none of the existing reviews are dedicated to the deployment of ML onto EO satellites for the purpose of on-board processing' (Section 1), but it does not report any systematic search protocol: no databases queried, no query strings, no inclusion/exclusion criteria, no date cutoff, and no screening process. Because the paper's contribution is entirely a synthesis of prior work, its coverage is the load-bearing claim. Without a methodology, the novelty claim is unverifiable and the selection of referenced works cannot be distinguished from a convenience sample. The admission in Section 4.3 that 'the author could not find a reference paper with such work done' further indicates that the literature search was informal. I recommend adding a methodology section that specifies the search strategy, screening criteria, and date range, and tempering the novelty claim to explicitly refer to the works the authors surveyed.
  2. [Section 4.1, Reference [55]] The quantitative claim that quantization reduces inference power consumption by up to 87% is supported solely by the authors' own previous work (Duggan et al. [55]). In a survey, a single self-cited result is a weak basis for a general conclusion about the effectiveness of quantization. The authors should cite independent evaluations of quantization on edge processors (for example, studies using TensorRT or TFLite) and should state the exact experimental conditions under which the 87% figure was obtained, including device, model architecture, bit-width, and measurement methodology. At minimum, the self-citation should be explicitly flagged as the authors' own result.
  3. [Sections 3 and 4] The survey would be considerably more useful if it included a comparative table summarizing the key characteristics of the reviewed on-board processing experiments and missions: hardware platform (Myriad 2, Jetson TX2/Nano/Orin, FPGA, etc.), model architecture, accuracy, power consumption, memory footprint, and whether the system was deployed in orbit or tested on the ground. Without such a synthesis, the paper reads as a narrative list of individual works, and the stated research question of which techniques are 'most effective' is never explicitly answered. A comparative table would also help substantiate the claim that the review is thorough rather than merely a selection of works.
minor comments (6)
  1. [Abstract] The phrase 'reduction in it's cost' contains a grammatical error; 'it's' should be 'its'. Similar possessive misuse appears elsewhere in the text.
  2. [Section 2.1.5] The sentence 'The global Small Satellite market size is projected to grow annually by over 20% to $6.6 by the end of 2024' is missing units after '$6.6'; it should read '$6.6 billion'.
  3. [Section 3.2] The term 'altitude control' should be 'attitude control', referring to the spacecraft's orientation subsystem.
  4. [Section 4.3] The anecdote about the Irish Meteorological Service and cloud cover over Ireland is out of place in a global survey; if the intent is to motivate cloud-filtering, a global cloud-cover statistic would be more appropriate.
  5. [Section 4.3] The sentence 'Unfortunately the author could not find a reference paper with such work done' uses the singular 'author' inconsistently with the elsewhere plural authorial voice; it should be 'the authors'.
  6. [References] Several references are incomplete and difficult to locate: [7], [16], [33], and [74] lack author names, full titles, or publication venues. The IEEE reference format should be applied consistently.

Circularity Check

0 steps flagged · score 0.0 of 10

No circularity: the survey makes no derived predictions, and its only self-citation is an external empirical reference, not a load-bearing premise.

full rationale

This paper is a literature survey rather than a derivation, so there is no chain of equations or fitted parameters whose output could reduce to its input. The central claims—that power, processing capability, memory, and radiation are the main constraints, and that quantization, lightweight architectures, accelerators, and on-board pre-processing mitigate them—are summaries of the cited external literature, with each mitigation claim attached to specific prior studies (e.g., CloudScout [66], FPGA benchmarking [71], Jetson evaluations [59]–[65]). The one self-citation, Duggan et al. [55], is used in Section 4.1 to report measured power reductions from quantization; that is an empirical result from the authors' own earlier conference paper, external to this survey, and the survey's recommendation does not depend solely on it because quantization is also supported by the general optimization literature it cites. The 'first dedicated review' novelty claim in Section 1 is not backed by a systematic search protocol, but this is a completeness/evidence weakness rather than circular reasoning: the claim is about the state of the literature, not derived from the survey's own definitions or outputs. No step in the paper equates a predicted quantity with an input by construction, and no load-bearing argument rests on a self-citation chain.

Assumptions & free parameters 0 free parameters · 2 assumptions · 0 invented entities

As a survey, the paper introduces no new parameters, entities, or derivations. Its conclusions rest on the trustworthiness and completeness of the cited literature, and on the assumption that the four selected constraints are the dominant ones. These are domain assumptions rather than formal axioms.

assumptions (2)
  • domain assumption The cited literature accurately reflects the state of the art in on-board AI image processing for Earth observation satellites.
    The survey's value depends on the accuracy and representativeness of the papers it cites, but it provides no independent verification.
  • domain assumption The four constraints (power, processing capability, memory, radiation) are the most significant barriers to deploying pre-trained ML models on-board EO satellites.
    This categorization is asserted in Sections 3 and 4 without a formal justification or comparison to other potential constraints.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Advancing Earth Observation: A Survey on AI-Powered Image Processing in Satellites." pith.science (2026). https://pith.science/paper/TOQJQNVQ

@misc{pith2026250112030,
  author       = {Pith},
  title        = {Pith review of: Advancing Earth Observation: A Survey on AI-Powered Image Processing in Satellites},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/TOQJQNVQ}},
  note         = {Machine review of arXiv:2501.12030}
}
read the original abstract

Advancements in technology and reduction in it's cost have led to a substantial growth in the quality & quantity of imagery captured by Earth Observation (EO) satellites. This has presented a challenge to the efficacy of the traditional workflow of transmitting this imagery to Earth for processing. An approach to addressing this issue is to use pre-trained artificial intelligence models to process images on-board the satellite, but this is difficult given the constraints within a satellite's environment. This paper provides an up-to-date and thorough review of research related to image processing on-board Earth observation satellites. The significant constraints are detailed along with the latest strategies to mitigate them.

Figures

Figures reproduced from arXiv: 2501.12030 by the authors.

Figure 1
Figure 1. LEO Satellite Block Diagram [12] 1999 [13] Puig-Suari and Bob Twiggs developed the Cube￾Sat specifications to promote the design of small satellites intended for low Earth orbit (LEO). This form factor, of 10 cm cubes weighing no more than 2kg, has become the de-facto standard for small EO satellites (smallsats) design. Satellites are defined by the number of cubes (e.g. 6U is 6 cubes stacked in a 2 cubes side-by-si… view at source ↗
Figure 2
Figure 2. Remote Sensing [14] there are 2 examples of external sources (1 - Sun, 2 - Earth). These are generally categorised as optical sensors operating at wavelengths typical for visible light(430–720 nm) or for infra-red(750–950 nm) and while they can get high resolution of small areas, they can only function during the day (i.e. need sunlight) and are affected by cloud and weather. Hyperspectral sensors collect informatio… view at source ↗
Figure 3
Figure 3. Power consumption during day with transmission & payload ( [37] [PITH_FULL_IMAGE:figures/full_fig_p005_3.png] view at source ↗
Figures from the paper (4 more)
Figure 4
Figure 4. Figure 4: Power consumption over 24 hours consumption of each subsystem over the full day (24 hours) is shown in [PITH_FULL_IMAGE:figures/full_fig_p006_4.png]
Figure 5
Figure 5. Figure 5: Comparison of models FLOPS to try and take advantage of this redundancy which broadly classify optimisation approaches under the following: • Layer decomposition - an approach to reduce the network computational complexity by reducing the size of individual CNN layers.…
Figure 6
Figure 6. Figure 6: Performance scaled by power comparison of onboard processors ( [56] [PITH_FULL_IMAGE:figures/full_fig_p009_6.png]
Figure 7
Figure 7. Figure 7: Model Memory Requirements Comparison Some model optimisation techniques (Layer decomposi￾tion, Knowledge Distillation,Pruning & Quantization ) were detailed in the previous power reduction section 4.1. Along with reducing the power required to run these models they can…

Discussion (0). Continue with ORCID to comment.

Forward citations

Cited by 1 Pith paper

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score. Full citation record

  1. SAT-Edge-Agent: Hardware-in-the-Loop Edge-Agent Orchestration for Onboard Satellite Intelligence

    cs.AI 2026-08 conditional novelty 5.0 of 10

    An edge-agent system on a COTS ARM SoC repeatedly completed two fixed FAIR1M detection workflows 20/20 times, with detector time only about 2.5% to 2.9% of full-agent latency.

Reference graph

Works this paper leans on

78 extracted references · 75 canonical work pages · cited by 1 Pith paper

  1. [55]

    & Giltinan, A

    Duggan, A., Scully, T., Smith, N. & Giltinan, A. Profiling Power Consumption for Deep Learning on Resource Limited Devices in International Conference on Innovative Techniques and Applications of Artificial Intelligence (2023), 129–141

    work page 2023

  2. [1]

    https://orbit.ing-now.com/

    2024. https://orbit.ing-now.com/

    work page 2024

  3. [2]

    Furano, G. et al. Towards the Use of Artificial Intelligence on the Edge in Space Systems: Challenges and Opportu- nities. IEEE Aerospace and Electronic Systems Magazine 35, 44–56 (2020)

    work page 2020

  4. [3]

    Future intelligent earth observing satellites in Earth Observing Systems VIII 5151 (2003), 1–8

    Zhou, G. Future intelligent earth observing satellites in Earth Observing Systems VIII 5151 (2003), 1–8

    work page 2003

  5. [4]

    & Stepanova, D

    Di Fraia, Z., Wischert, D. & Stepanova, D. Machine Learn- ing in Earth Observation Operations: A review Pablo Mi- ralles, Antonio Fulvio Scannapieco, Nitya Jagadam, Pre- rna Baranwal, Bhavin Faldu, Ruchita Abhang, Sahil Bha- tia, Sebastien Bonnart, Ishita Bhatnagar, Beenish Batul, Pallavi Prasad, H ´ector Ortega-Gonz ´alez, Harrish Joseph, Harshal More, S...

    work page 2021

  6. [5]

    Miralles, P . et al. A critical review on the state-of-the-art and future prospects of Machine Learning for Earth Observa- tion Operations. Advances in Space Research (2023)

    work page 2023

  7. [6]

    Zhang, B. et al. Progress and challenges in intelligent re- mote sensing satellite systems. IEEE Journal of Selected Top- ics in Applied Earth Observations and Remote Sensing 15, 1814–1822 (2022)

    work page 2022

  8. [7]

    https : / / api

    Towards Space Edge Computing and Onboard AI for Real-Time Teleoperations in (). https : / / api . semanticscholar . org / CorpusID:268615001

    work page

Show all 78 references
  1. [8]

    Thangavel, K. et al. Artificial intelligence for trusted au- tonomous satellite operations. Progress in Aerospace Sciences 144, 100960 (2024)

    2024

  2. [9]

    Earth satellite https://www.britannica.com/technology/ Earth-satellite

  3. [10]

    nasa.gov/features/OrbitsCatalog

    Catalog of Earth Satellite Orbits https://earthobservatory. nasa.gov/features/OrbitsCatalog

  4. [11]

    Launch costs to low Earth orbit, 1980-2100 — Future Timeline — Data & Trends — Future Predictions 2020

    Fox, W. Launch costs to low Earth orbit, 1980-2100 — Future Timeline — Data & Trends — Future Predictions 2020. https: //www.futuretimeline.net/data-trends/6.htm

    1980

  5. [12]

    & El-Emam, E

    El-Bayoumi, A., Salem, M., Khalil, A. & El-Emam, E. A new Checkout-and-Testing-Equipment (CTE) for a satellite Telemetry using LabVIEW in 2015 IEEE Aerospace Conference (2015), 1– 9

    2015

  6. [13]

    & Twiggs, R

    Heidt, H., Puig-Suari, J., Moore, A., Nakasuka, S. & Twiggs, R. CubeSat: A new generation of picosatellite for education and industry low-cost space experimentation (2000)

    2000

  7. [14]

    omnisci.com/technical-glossary/remote-sensing

    What is Remote Sensing? Definition and FAQs https://www. omnisci.com/technical-glossary/remote-sensing

  8. [15]

    Space Data: The Final Analytics Frontier

    Data, P . Space Data: The Final Analytics Frontier. Intel. https : / / www. intel . co . uk / content / www / uk / en / it - management / cloud - analytic - hub / big - data - from - satellites.html (2017)

    2017

  9. [16]

    https : / / www

    2024. https : / / www. thebusinessresearchcompany. com / report/small-satellite-global-market-report

    2024

  10. [17]

    Esch, T. et al. Exploiting big earth data from space–first experiences with the timescan processing chain. Big Earth Data 2, 36–55 (2018)

    2018

  11. [18]

    Spotting satellites

    Beall, A. Spotting satellites. New Scientist 247, 51 (2020)

    2020

  12. [19]

    Earth Sciences

    satellite ground stations - everything you ever wanted to know and more. Earth Sciences. https : / / www. essearth . com/satellite-ground-stations/ (2020)

    2020

  13. [20]

    & Hagmanns, F.-J

    Wertz, P ., Hespeler, B., Kiessling, M. & Hagmanns, F.-J. Next generation high data rate downlink subsystems based on a flexible APSK modulator applying SCCC encoding in 2016 International Workshop on Tracking, Telemetry and Command Systems for Space Applications (TTC) (2016), 1–7

    2016

  14. [21]

    W., Zhai, G., Wang, W

    Shi, B., Leong, S. W., Zhai, G., Wang, W. & Luo, B. Develop- ment of Ka-band BUC with wideband linearizer for high speed satellite communications in GLOBECOM 2017-2017 IEEE Global Communications Conference (2017), 1–6

    2017

  15. [22]

    & Rovatti, M

    Furano, G., Tavoularis, A. & Rovatti, M. AI in space: Ap- plications examples and challenges in 2020 IEEE International Symposium on Defect and Fault Tolerance in VLSI and Nan- otechnology Systems (DFT) (2020), 1–6

    2020

  16. [23]

    Sowmya, D., Shenoy, P . D. & Venugopal, K. Remote sensing satellite image processing techniques for image classifica- tion: a comprehensive survey. International Journal of Com- puter Applications 161, 24–37 (2017)

    2017

  17. [24]

    & Golla, S

    Abburu, S. & Golla, S. B. Satellite image classification meth- ods and techniques A review. International journal of com- puter applications 119 (2015)

    2015

  18. [25]

    & Rathee, N

    Babbar, J. & Rathee, N. Satellite Image Analysis: A Review in 2019 IEEE International Conference on Electrical, Computer and Communication Technologies (ICECCT) (2019), 1–6

    2019

  19. [26]

    & Belangour, A

    Ouchra, H. & Belangour, A. Satellite image classification methods and techniques: A survey in 2021 IEEE International Conference on Imaging Systems and Techniques (IST) (2021), 1–6

    2021

  20. [27]

    & Hinton, G

    Krizhevsky, A., Sutskever, I. & Hinton, G. E. Imagenet clas- sification with deep convolutional neural networks. Ad- vances in neural information processing systems 25, 1097–1105 (2012)

    2012

  21. [28]

    & Guo, Y

    Hao, S., Zhou, Y. & Guo, Y. A brief survey on seman- tic segmentation with deep learning. Neurocomputing 406, 302–321 (2020)

    2020

  22. [29]

    & Sun, J

    He, K., Zhang, X., Ren, S. & Sun, J. Deep residual learning for image recognition in Proceedings of the IEEE conference on computer vision and pattern recognition (2015), 770–778

    2015

  23. [30]

    L., Rochester, N

    McCarthy, J., Minsky, M. L., Rochester, N. & Shannon, C. E. A proposal for the dartmouth summer research project on artificial intelligence, august 31, 1955.AI magazine 27, 12–12 (2006)

    2006

  24. [31]

    com / product/state- of- iot- spring- 2024//

    State of IoT – Spring 2024 https : / / iot - analytics . com / product/state- of- iot- spring- 2024//. Accessed: 2024-03- 25

    2024

  25. [32]

    & Shin, K

    Lee, J., Kim, E. & Shin, K. G. Design and management of satellite power systems in 2013 IEEE 34th Real-Time Systems Symposium (2013), 97–106

    2013

  26. [33]

    https : / / www

    2023. https : / / www . ucsusa . org / resources / satellite - database

    2023

  27. [34]

    Lentaris, G. et al. High-performance embedded computing in space: Evaluation of platforms for vision-based navi- gation. Journal of Aerospace Information Systems 15, 178–192 (2018). 12

    2018

  28. [35]

    H., Fraeman, M

    Maurer, R. H., Fraeman, M. E., Martin, M. N. & Roth, D. R. Harsh environments: space radiation. Johns Hopkins APL technical digest 28, 17 (2008)

    2008

  29. [36]

    & Apruzzese, G

    Lange, K., Fontana, F., Rossi, F., Varile, M. & Apruzzese, G. Machine Learning in Space: Surveying the Robust- ness of on-board ML models to Radiation. arXiv preprint arXiv:2405.02642 (2024)

    2024 arXiv

  30. [37]

    Dahbi, S. et al. Power budget analysis for a LEO polar orbiting nano-satellite in 2017 International Conference on Advanced Technologies for Signal and Image Processing (ATSIP) (2017), 1–6

    2017

  31. [38]

    Gill, S. S. et al. Modern computing: Vision and challenges. Telematics and Informatics Reports, 100116 (2024)

    2024

  32. [39]

    & Zhang, Z

    Sze, V ., Chen, Y.-H., Emer, J., Suleiman, A. & Zhang, Z. Hardware for machine learning: Challenges and opportunitiesin 2017 IEEE custom integrated circuits conference (CICC)(2017), 1–8

    2017

  33. [40]

    Reuther, A. et al. AI and ML accelerator survey and trends in 2022 IEEE High Performance Extreme Computing Conference (HPEC) (2022), 1–10

    2022

  34. [41]

    Ortiz, F. et al. Onboard processing in satellite communica- tions using ai accelerators. Aerospace 10, 101 (2023)

    2023

  35. [42]

    A., Duman, B

    S ¨uzen, A. A., Duman, B. & S ¸en, B. Benchmark analysis of jetson tx2, jetson nano and raspberry pi using deep-cnn in 2020 International Congress on Human-Computer Interaction, Optimization and Robotic Applications (HORA) (2020), 1–5

    2020

  36. [43]

    & Zisserman, A

    Simonyan, K. & Zisserman, A. Very deep convolutional networks for large-scale image recognition. arXiv preprint arXiv:1409.1556 (2014)

    2014 arXiv

  37. [44]

    & Alemi, A

    Szegedy, C., Ioffe, S., Vanhoucke, V . & Alemi, A. A. Inception-v4, inception-resnet and the impact of residual connec- tions on learning in Thirty-first AAAI conference on artificial intelligence (2016)

    2016

  38. [45]

    The vanishing gradient problem during learning recurrent neural nets and problem solutions

    Hochreiter, S. The vanishing gradient problem during learning recurrent neural nets and problem solutions. In- ternational Journal of Uncertainty, Fuzziness and Knowledge- Based Systems 6, 107–116 (1998)

    1998

  39. [46]

    & Sun, J

    He, K., Zhang, X., Ren, S. & Sun, J. Delving deep into recti- fiers: Surpassing human-level performance on imagenet classi- fication in Proceedings of the IEEE international conference on computer vision (2015), 1026–1034

    2015

  40. [47]

    Russakovsky, O. et al. Imagenet large scale visual recogni- tion challenge. International journal of computer vision 115, 211–252 (2015)

    2015

  41. [48]

    & Sun, G

    Hu, J., Shen, L. & Sun, G. Squeeze-and-excitation networks in Proceedings of the IEEE conference on computer vision and pattern recognition (2018), 7132–7141

    2018

  42. [49]

    Iandola, F. N. et al. SqueezeNet: AlexNet-level accuracy with 50x fewer parameters and¡ 0.5 MB model size. arXiv preprint arXiv:1602.07360 (2016)

    2016 arXiv

  43. [50]

    Tan, M. et al. Mnasnet: Platform-aware neural architecture search for mobile in Proceedings of the IEEE/CVF conference on computer vision and pattern recognition (2019), 2820–2828

    2019

  44. [51]

    Howard, A. G. et al. Mobilenets: Efficient convolutional neural networks for mobile vision applications. arXiv preprint arXiv:1704.04861 (2017)

    2017 arXiv

  45. [52]

    & Sun, J

    Zhang, X., Zhou, X., Lin, M. & Sun, J. Shufflenet: An ex- tremely efficient convolutional neural network for mobile devices in Proceedings of the IEEE conference on computer vision and pattern recognition (2018), 6848–6856

    2018

  46. [53]

    J., Li, X

    Wang, R. J., Li, X. & Ling, C. X. Pelee: A real-time object detection system on mobile devices. Advances in neural in- formation processing systems 31 (2018)

    2018

  47. [54]

    V ´estias, M. P . A survey of convolutional neural networks on edge with reconfigurable computing. Algorithms 12, 154 (2019)

    2019

  48. [56]

    George, A. D. & Wilson, C. M. Onboard processing with hybrid and reconfigurable computing on small satellites. Proceedings of the IEEE 106, 458–470 (2018)

    2018

  49. [57]

    G., Weisz, G., French, M., Flatley, T

    Schmidt, A. G., Weisz, G., French, M., Flatley, T. & Vil- lalpando, C. Y. SpaceCubeX: A framework for evaluating hy- brid multi-core CPU/FPGA/DSP architectures in 2017 IEEE Aerospace Conference (2017), 1–10

    2017

  50. [58]

    Geist, A. et al. SpaceCube v3. 0 NASA next-generation high-performance processor for science applications (2019)

    2019

  51. [59]

    Buonaiuto, N. et al. Satellite identification imaging for small satellites using NVIDIA (2017)

    2017

  52. [60]

    P ., Michaels, A

    Arechiga, A. P ., Michaels, A. J. & Black, J. T. Onboard image processing for small satellitesin NAECON 2018-IEEE National Aerospace and Electronics Conference (2018), 234–240

    2018

  53. [61]

    Manning, J. et al. Machine-learning space applications on smallsat platforms with tensorflow (2018)

    2018

  54. [62]

    & Zhou, Y

    Yao, Y., Jiang, Z., Zhang, H. & Zhou, Y. On-board ship detection in micro-nano satellite based on deep learning and COTS component. Remote Sensing 11, 762 (2019)

    2019

  55. [63]

    A Survey on optimized implementation of deep learning models on the NVIDIA Jetson platform

    Mittal, S. A Survey on optimized implementation of deep learning models on the NVIDIA Jetson platform. Journal of Systems Architecture 97, 428–442 (2019)

    2019

  56. [64]

    S., Tiwari, N

    Slater, W. S., Tiwari, N. P ., Lovelly, T. M. & Mee, J. K. Total ionizing dose radiation testing of NVIDIA Jetson nano GPUsin 2020 IEEE High Performance Extreme Computing Conference (HPEC) (2020), 1–3

    2020

  57. [65]

    Hern ´andez-G´omez, J. et al. Conceptual low-cost on-board high performance computing in CubeSat nanosatellites for pattern recognition in Earth’s remote sensing in Proceedings of the 1st International Con 13 (2019), 114–122

    2019

  58. [66]

    Giuffrida, G. et al. CloudScout: A deep neural network for on-board cloud detection on hyperspectral images. Remote Sensing 12, 2205 (2020)

    2020

  59. [67]

    phi-sat https://www.esa.int/Applications/Observing the Earth/Ph-sat/

  60. [68]

    Deniz, O. et al. Eyes of things. Sensors 17, 1173 (2017)

    2017

  61. [69]

    & Zong, Z

    Li, D., Chen, X., Becchi, M. & Zong, Z. Evaluating the energy efficiency of deep convolutional neural networks on CPUs and GPUs in 2016 IEEE international conferences on big data and cloud computing (BDCloud), social computing and network- ing (SocialCom), sustainable computin...

    2016

  62. [70]

    Giuffrida, G. et al. The Φ-Sat-1 mission: The first on-board deep neural network demonstrator for satellite earth obser- vation. IEEE Transactions on Geoscience and Remote Sensing 60, 1–14 (2021)

    2021

  63. [71]

    Rapuano, E. et al. An FPGA-Based Hardware Accelera- tor for CNNs Inference on Board Satellites: Benchmarking with Myriad 2-Based Solution for the CloudScout Case Study. Remote Sensing 13, 1518 (2021)

    2021

  64. [72]

    & Brown, D

    Barnell, M., Raymond, C., Smiley, S., Isereau, D. & Brown, D. Ultra Low-Power Deep Learning Applications at the Edge with Jetson Orin AGX Hardware in 2022 IEEE High Perfor- mance Extreme Computing Conference (HPEC) (2022), 1–4. 13

    2022

  65. [73]

    O., Alarcia, R

    Rad, I. O., Alarcia, R. M. G., Dengler, S., Golkar, A. & Man- fletti, C. Preliminary Evaluation of Commercial Off-The- Shelf GPUs for Machine Learning Applications in Space (2023)

    2023

  66. [74]

    https : / / unibap

    2023. https : / / unibap . com / solutions / spacecloud - hardware/ix10/

    2023

  67. [75]

    https://kplabs.space/antelope/

    2023. https://kplabs.space/antelope/

    2023

  68. [76]

    & Cho, M

    Maskey, A. & Cho, M. CubeSatNet: Ultralight Convolu- tional Neural Network designed for on-orbit binary image classification on a 1U CubeSat. Engineering Applications of Artificial Intelligence 96, 103952 (2020)

    2020

  69. [77]

    https://birds3.birds-project.com/

    2021. https://birds3.birds-project.com/

    2021

  70. [78]

    & Song, J

    Zhang, Z., Xu, G. & Song, J. Cubesat cloud detection based on JPEG2000 compression and deep learning. Advances in Mechanical Engineering 10, 1687814018808178 (2018)

    2018

Pith tools

Reviewed August 10, 2026 · model on record in the stance chip above.