Recognition and Classification of Ship Noise in Underwater Soundscape Based on Yet Another Mobile Network and Unmanned Surface Vehicle
Journal
2025 IEEE Underwater Technology, UT 2025
Part Of
2025 IEEE Underwater Technology, UT 2025
Start Page
1
End Page
5
Date Issued
2025-03-02
Author(s)
Abstract
This paper proposes using Yet another mobile network (YAMNet) to detect engine noise from ships within the underwater soundscape. This novel approach combines deep learning with sound classification techniques to accurately identify and categorize engine noises from various vessels in the marine environment. It holds significant potential for marine ecological protection, ship management, and noise pollution control applications. YAMNet is a deep learning model based on the MobileNet architecture, designed for audio event detection and classification. MobileNet is a lightweight convolutional neural network (CNN) known for its high computational efficiency and compactness, making it suitable for execution on resource-constrained Internet of Things (IoT) devices through edge computing. YAMNet was originally designed to effectively process audio data from various environments by converting sound signals into spectrograms, which are then processed through the deep learning network for feature extraction and classification. This enables YAMNet to quickly and accurately recognize various audio events, such as natural sounds, mechanical noises, and human activities.
Event(s)
2025 IEEE Underwater Technology, UT 2025, Taipei, 2 March 2025 through 5 March 2025. Code 208171
Subjects
Convolutional Neural Network
ships noise
Underwater sound scape
Yet Another Mobile Network
SDGs
Publisher
IEEE
Type
conference paper