Digitala Vetenskapliga Arkivet

Change search
CiteExportLink to record
Permanent link

Direct link
Cite
Citation style
  • apa
  • ieee
  • modern-language-association-8th-edition
  • vancouver
  • Other style
More styles
Language
  • de-DE
  • en-GB
  • en-US
  • fi-FI
  • nn-NO
  • nn-NB
  • sv-SE
  • Other locale
More languages
Output format
  • html
  • text
  • asciidoc
  • rtf
A Systematic Comparison of Data Augmentation Strategies for Vessel Classification and Trajectory Prediction
Linnaeus University, Faculty of Technology, Department of computer science and media technology (CM).
2025 (English)Independent thesis Advanced level (degree of Master (One Year)), 10 credits / 15 HE creditsStudent thesis
Abstract [en]

Maritime transportation plays a vital role in global trade; however, the sparse, noisy, and incomplete nature of data from the maritime Automatic Identification System (AIS) presents challenges for predictive analytics and hinders research progress in this field. This project investigates the effectiveness of data augmentation (DA) techniques in improving deep learning (DL) models for two tasks: vessel type classification and trajectory prediction. To address the lack of systematic comparisons among DA methods, this study proposes a modular pipeline that integrates AIS data preprocessing, synthetic data generation utilizing three levels of DA techniques, and subsequent statistical analysis using the Wilcoxon rank-sum and Kruskal-Wallis tests. The DA techniques range from a simple k-Means approach to an advanced method employing Variational Autoencoders, including a domain-specific approach that integrates geographical noise. Controlled experiments demonstrated that DA improved the DL model’s predictive capability and indicated significant differences among the techniques, with domain-specific methods showing particular promise. These findings assist in selecting appropriate DA techniques for specific tasks, and the pipeline lays a foundation for practical applications that can be flexibly adapted to other tasks or DA methods.

Place, publisher, year, edition, pages
2025. , p. 48
Keywords [en]
data augmentation, deep learning, AIS data, maritime trajectories
National Category
Electrical Engineering, Electronic Engineering, Information Engineering
Identifiers
URN: urn:nbn:se:lnu:diva-137321OAI: oai:DiVA.org:lnu-137321DiVA, id: diva2:1946814
Subject / course
Computer Science
Supervisors
Examiners
Available from: 2025-04-09 Created: 2025-03-24 Last updated: 2025-04-09Bibliographically approved

Open Access in DiVA

fulltext(2605 kB)68 downloads
File information
File name FULLTEXT01.pdfFile size 2605 kBChecksum SHA-512
8bd1d8c8543bd65233f24d0acc44e6f3755b77dd6929498ef8b4276b0fb3c0b0a20c60cdaeae8421fe9dacbe7acbf20460e98e3ea794936a28812ca67f1b9404
Type fulltextMimetype application/pdf

Search in DiVA

By author/editor
Keck, Fabian
By organisation
Department of computer science and media technology (CM)
Electrical Engineering, Electronic Engineering, Information Engineering

Search outside of DiVA

GoogleGoogle Scholar
Total: 69 downloads
The number of downloads is the sum of all downloads of full texts. It may include eg previous versions that are now no longer available

urn-nbn

Altmetric score

urn-nbn
Total: 221 hits
CiteExportLink to record
Permanent link

Direct link
Cite
Citation style
  • apa
  • ieee
  • modern-language-association-8th-edition
  • vancouver
  • Other style
More styles
Language
  • de-DE
  • en-GB
  • en-US
  • fi-FI
  • nn-NO
  • nn-NB
  • sv-SE
  • Other locale
More languages
Output format
  • html
  • text
  • asciidoc
  • rtf