Lightweight Sonar Target Detection Model for Unmanned Underwater Vehicle via Multiscale Feature Fusion and Acoustic Perception
IEEE Sensors Journal · 15.10.2025 · 25 (20) · 38230-38244
DOI: 10.1109/jsen.2025.3597642 ↗Kurz gefasst
Fachartikel in der Zeitschrift IEEE Sensors Journal, erschienen am 15.10.2025. Verfasst von Kunde Yang und 5 weiteren Autor:innen. Die Arbeit wurde bisher 5-mal zitiert (Crossref) und nennt 36 Literaturangaben. Gefördert von: Key Program of the National Natural Science Foundation of China, National Natural Science Foundation of China, National Key Research and Development Program of China.
Abstract (Originalsprache)
To address the urgent need for efficient and accurate perceptual capabilities in unmanned underwater vehicle (UUV)-based information countermeasure systems, this article proposes a novel lightweight and edge-aware detection framework for forward-looking sonar images. Despite the advances in underwater target detection, the existing models are often hindered by high computational costs, suboptimal boundary localization, and the lack of specialized sonar datasets. In this work, we propose Lightweight-YOLOv8, an object detection model tailored for sonar applications, and construct a dedicated underwater sonar image (USI) dataset to support experimental validation. Specifically, we design a lightweight C2F (LWC2F) module integrating inverted residual blocks to achieve a balance between detection accuracy and inference speed. To improve target representation, we propose a multiscale and multireceptive field feature fusion strategy. Furthermore, an edge-aware CIoU (EACIoU) loss function is introduced by incorporating Canny edge detection to enhance boundary precision in sonar scenes. Experimental results on the USI dataset show that our model achieves a precision of 98.45%, surpassing YOLOv7 and YOLOv8 by 9.73% and 4.99%, respectively. Ablation studies confirm the contribution of each component. The proposed framework offers a promising solution for distributed, real-time acoustic perception in intelligent UUV missions.
Förderung
- Key Program of the National Natural Science Foundation of China
- National Natural Science Foundation of China
- National Key Research and Development Program of China
- Northwestern Polytechnical University Fund
- Joint Training Fund Project of Hanjiang National Laboratory
Zitieren
APA
Yang, K., Liu, F., Li, Z., Li, H., Chen, F., & Hao, W. (2025). Lightweight Sonar Target Detection Model for Unmanned Underwater Vehicle via Multiscale Feature Fusion and Acoustic Perception. IEEE Sensors Journal, 25(20), 38230-38244. https://doi.org/10.1109/jsen.2025.3597642
Vancouver
Yang K, Liu F, Li Z, Li H, Chen F, Hao W. Lightweight Sonar Target Detection Model for Unmanned Underwater Vehicle via Multiscale Feature Fusion and Acoustic Perception. IEEE Sensors Journal. 2025;25(20):38230-38244. doi:10.1109/jsen.2025.3597642
BibTeX
@article{Yang2025,
author = {Yang, Kunde and Liu, Feng and Li, Zipeng and Li, Hui and Chen, Fuhu and Hao, Wang},
title = {Lightweight Sonar Target Detection Model for Unmanned Underwater Vehicle via Multiscale Feature Fusion and Acoustic Perception},
journal = {IEEE Sensors Journal},
year = {2025},
volume = {25},
number = {20},
pages = {38230--38244},
doi = {10.1109/jsen.2025.3597642}
}
EUVIMED