Skip to main navigation Skip to search Skip to main content

Research on Ship Target Recognition Based on Optimized VMD and Transformer

  • Jiawen Yang
  • , Qiao Hu
  • , Zirui Wang
  • , Hongbo Wei
  • , Enhui Yang
  • , Guangyu Jiang
  • Xi'an Jiaotong University

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

Abstract

Underwater acoustic target recognition faces challenges due to low signal-to-noise ratio (SNR) in ship- radiated noise and limited dataset availability. To address these issues, this paper proposes an integrated approach combining variational mode decomposition (VMD) optimized by the grey wolf optimizer (GWO) for feature enhancement, Wasserstein generative adversarial networks (WGAN) with continuous wavelet transform for data augmentation, and a Swin-Transformer network for classification. Experimental results on the ShipsEar dataset demonstrate 96% recognition accuracy across five ship classes, validating the effectiveness of the proposed method for high-precision underwater target recognition.

Original languageEnglish
Title of host publicationProceedings - 2025 5th International Conference on Robotics, Automation and Intelligent Control, ICRAIC 2025
PublisherInstitute of Electrical and Electronics Engineers Inc.
ISBN (Electronic)9798331567446
DOIs
StatePublished - 2025
Event5th International Conference on Robotics, Automation and Intelligent Control, ICRAIC 2025 - Chengdu, China
Duration: 31 Oct 20252 Nov 2025

Publication series

NameProceedings - 2025 5th International Conference on Robotics, Automation and Intelligent Control, ICRAIC 2025

Conference

Conference5th International Conference on Robotics, Automation and Intelligent Control, ICRAIC 2025
Country/TerritoryChina
CityChengdu
Period31/10/252/11/25

Keywords

  • Generative Adversarial Network
  • Self-Attention
  • Ship Target
  • Variational Mode Decomposition

Fingerprint

Dive into the research topics of 'Research on Ship Target Recognition Based on Optimized VMD and Transformer'. Together they form a unique fingerprint.

Cite this