Student or educator? Sign up with your school email and get 6 months of Pro free.
CoChat

The DOI matches its Crossref record; 64 of the 71 references it cites were found in a registry or an index; 16 of the 55 sources its claims cite were checked for support.

Space-Based Global Maritime Surveillance. Part II: Artificial Intelligence and Data Fusion Techniques

Open access

Abstract

Maritime surveillance (MS) is of paramount importance for search and rescue operations, fishery monitoring, pollution control, law enforcement, migration monitoring, and national security policies. Since ground-based radars and automatic identification system (AIS) do not always provide a comprehensive and seamless coverage of the entire maritime domain, the use of space-based sensors is crucial to complement them. We reviewed space-based technologies for MS in the first part of this work, titled Space-based Global Maritime Surveillance. Part I: Satellite Technologies. However, MS systems combining multiple terrestrial and space-based sensors with additional information sources require dedicated artificial intelligence and data fusion techniques for processing raw satellite images and fusing heterogeneous information. The second part of our work focuses on some recent promising artificial intelligence and data fusion techniques for MS using space-based sensors.

Overview

CoChat Summary

TL;DR

This review surveys artificial intelligence, deep learning, and Bayesian/data-fusion techniques for space-based maritime surveillance, covering satellite image segmentation/classification, knowledge discovery from Sat-AIS, anomaly detection, and sum-product-algorithm-based multitarget tracking. It illustrates these methods with use cases on VHR/HR/Sentinel-2/COSMO-SkyMed imagery and AIS/SAR fusion, including an MTG extraction and a Malta-area SAR/AIS tracking scenario.

Key findings

  • DNN use cases include Mask R-CNN trained on about 200,000 VHR images (≈1.5 m resolution, 768×768 pixels) from GeoEye, SPOT, Pleiades, and Black Sky for ship detection/feature extraction; U-Net on a synthetic HR dataset obtained by degrading 10,000 VHR SPOT images; ResNet34 for stern/bow classification on Sentinel-2 clipped ships; and an AIS-based vessel classifier with 2 hidden layers (200 and 100 neurons) and 14 output categories.
  • Historical Sat-AIS knowledge discovery via maritime traffic graphs (MTGs) is demonstrated with 19,009 AIS tracks, 162,662 detected waypoints, 2,286 waypoint clusters, and a pruned graph of 99 nodes and 401 edges; the method was validated on real AIS datasets from the Iberian Coast and English Channel and in four MARISA operational trials.
  • Maritime anomaly detection is formulated using the Ornstein-Uhlenbeck mean-reverting model with generalized likelihood ratio testing and joint Bayesian detection/tracking via hybrid Bernoulli filtering; it can detect deviations from standard routes and “dark” deviations during surveillance gaps, and was applied to a real-world rendezvous detection scenario.
  • SPA-based multisensor MTT is presented as flexible and scalable for heterogeneous fusion, Sat-AIS integration, and class-aided tracking; a use case off Malta from 8 September 2018 at 16:30 to 12 September 2018 at 16:47 fused COSMO-SkyMed SAR measurements with AIS messages (AIS interval at least 5 hours) and identified a single target, the vehicles carrier GAIA LEADER, with real and estimated trajectories.

Methods

  • Design: review/overview paper, companion to Part I, synthesizing AI/DNN and Bayesian/data-fusion techniques for maritime surveillance using space-based sensors.
  • Data/sample: satellite images (VHR ≈1.5 m, HR ≈10 m, Sentinel-2, COSMO-SkyMed SAR), historical Sat-AIS data, AIS messages, and SAR measurements; use cases include training datasets of about 200,000 VHR images and 10,000 degraded VHR images.
  • Main analysis: descriptive review and illustrative use cases rather than a controlled comparative experiment; quantitative results are limited mainly to dataset/graph counts and the Malta SAR/AIS tracking example.

Limitations

  • The paper does not state an explicit limitations section. It notes Sat-AIS data challenges: erroneous, incomplete, intermittent, or counterfeit information; irregular sampling; coverage gaps from minutes to many hours; and intentional AIS disablement.
  • It states limitations of existing MTT algorithms: fusion of heterogeneous information sources is not straightforward, they do not adapt to time-varying model parameters, and their complexity usually does not scale well with relevant system parameters such as the number of sensors.
  • Evident from the text: as a review, the use cases are illustrative and not comprehensive; most DNN use cases report no accuracy or performance metrics, and the second HR segmentation use case relies on synthetic data because a real HR dataset was unavailable (the latter is stated in the paper).

Model-written summary. Check it against the paper before relying on it.

Integrity signals

Registry

Crossref · 10.1109/maes.2021.3070884

Cites

64 of 64 references verified0 citation errors · 0 not checkable

Statistics

Nothing to recompute.

Verification

Claims the paper makes

9 Valid · 8 Not valid · 19 Not checked · 41 Own results and interpretations

34 more claims · 7 with their cited sources checked

Create a free account to see all 77 claims and to verify the 2 references still open.

See all claims — free account
  • Probably not a substantial claim. 27
  • It may state two separate claims as one. 16
  • Borderline: one of the tests for a substantial claim is unclear. 1

Create a free account to see these claims.

Citations

Cites

Show all 71 references

Research works printed but not linked: 1

  1. adelson, 1984, Pyramid methods in image processing, RCA Eng, 29, 33

Cited by

See all 44

Metadata

Authors
Soldi, G., Gaglione, D., Forti, N., Millefiori, L. M., Braca, P., Carniel, S., Simone, A. D., Iodice, A., Riccio, D., Daffina, F. C., Quattrociocchi, D., Bottini, G., Willett, P., & Farina, A.
Affiliations
  • NATO Centre for Maritime Research and Experimentation (IT)
  • University of Naples Federico II (IT)
  • e GEOS (Italy) (IT)
  • Food and Agriculture Organization of the United Nations (IT)
  • University of Connecticut (US)
Venue
IEEE Aerospace and Electronic Systems Magazine
Published
2021-09-01
Volume
36
Issue
9
Pages
30-42
ISSN
  • 0885-8985 (ISSN-L)
  • 1557-959X
OpenAlex
W3110443977
Cited by
44
References
71
Citations by year
2026: 6 · 2025: 7 · 2024: 11 · 2023: 6 · 2022: 7 · 2021: 7
FWCI
4.11
Type
journal article
Language
en
Open access
Locations
  • IEEE Aerospace and Electronic Systems Magazine (publishedVersion)
  • arXiv (Cornell University) (submittedVersion)
  • CINECA IRIS Institutial research information system (University of Pisa) (submittedVersion)
  • Florence Research (University of Florence) (submittedVersion)
  • arXiv (Cornell University)
Funders
  • NATO Allied Command Transformation (SAC000A08)
  • NIUVT
  • Air Force Office of Scientific Research (FA9500-18-1-0463)
Topics
  • Physical Sciences › Engineering › Ocean Engineering › Maritime Navigation and Safety
  • Physical Sciences › Engineering › Media Technology › Advanced Image Fusion Techniques
  • Physical Sciences › Computer Science › Artificial Intelligence › Target Tracking and Data Fusion in Sensor Networks
Keywords
maritime surveillance, satellite sensors, satellite imagery, artificial intelligence, data fusion, heterogeneous data fusion, search and rescue, fisheries monitoring, pollution monitoring, migration monitoring, national security, Automatic Identification System

About these checks

Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.

The claim verdicts this report counts are experimental: their accuracy is not yet measured.

A language model drew the claims from the paper on 2026-10-10, under version 3 of the claim definition (model deepseek-v4.1-flash, prompt extract-v3).

A second model labelled each claim on 2026-10-10 (model jev-1.13-20260917, definition revision gold-guideline-v0.2).

Cited sources are checked with a language model’s help (qwen3.8-omni-flash and gpt-6-sol).

Statistics and arithmetic are recomputed, not judged.

Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.

What counts as a claim ›How we measure confidence ›How we check this ›

Are you an author, or do you see a problem with this page? Request a review

How to cite

Cite CoChat for the checks on this page, and the paper for its own findings.

This page
CoChat. (2026). Citation check: Space-Based Global Maritime Surveillance. Part II: Artificial Intelligence and Data Fusion Techniques. https://app.cochat.ai/papers/10.1109/maes.2021.3070884
The paper
Soldi, G., Gaglione, D., Forti, N., Millefiori, L. M., Braca, P., Carniel, S., Simone, A. D., Iodice, A., Riccio, D., Daffina, F. C., Quattrociocchi, D., Bottini, G., Willett, P., & Farina, A. (2021). Space-Based Global Maritime Surveillance. Part II: Artificial Intelligence and Data Fusion Techniques. IEEE Aerospace and Electronic Systems Magazine. https://doi.org/10.1109/maes.2021.3070884

Data sources

Bibliographic data: Semantic Scholar, OpenAlex, Crossref.

Contains information from the Semantic Scholar Academic Graph Datasets, made available under the ODC Attribution License (ODC-BY 1.0).

Semantic Scholar’s data is described in: Kinney, R., et al. (2023). The Semantic Scholar Open Data Platform. arXiv:2301.10140.