Sea area signal investigation method and device based on deep learning

By embedding a lightweight LeNet-5 hierarchical convolutional neural network model on an STM32H743 microcontroller, the problem of real-time identification of underwater acoustic signals in complex marine environments by underwater reconnaissance systems is solved, achieving efficient and accurate automatic identification and classification of underwater acoustic signals, which is suitable for front-end platforms such as buoys and submersibles.

CN121996983APending Publication Date: 2026-05-08HARBIN ENG UNIV
0 Cites 0 Cited by

Patent Information

Application Number
CN202610205431.4
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2026-02-12
Publication Date
2026-05-08

AI Technical Summary

Technical Problem

Existing underwater reconnaissance systems cannot automatically identify underwater acoustic signals in a long-term, real-time and accurate manner. Traditional neural networks have low computational efficiency and insufficient robustness in noisy and complex marine environments, and cannot provide real-time intelligence support for underwater non-cooperative target reconnaissance.

Method used

A lightweight LeNet-5 hierarchical convolutional neural network model is adopted and embedded into an STM32H743 microcontroller using the X-Cube-AI tool to achieve automatic identification and classification of underwater acoustic signals. This includes hierarchical processing of LA, LB, LC, and LD subnetworks. Combined with depthwise separable convolution, global average pooling, and SE attention module, the computational complexity and number of parameters are reduced.

Benefits of technology

It achieves long-term, real-time, and accurate automatic identification of underwater acoustic signals, improves signal processing efficiency and system real-time response capability, has good robustness and identification accuracy, and is suitable for running on resource-constrained embedded devices.

✦ Generated by Eureka AI based on patent content.
Patent Text Reader

Abstract

The invention belongs to the technical field of underwater acoustic communication reconnaissance and artificial intelligence edge computing, and discloses a sea area signal reconnaissance method and device based on deep learning. According to the method, a constructed lightweight LeNet-5 hierarchical convolutional neural network model is embedded and deployed into an STM32H743 microcontroller through an X-Cube-AI tool to process underwater acoustic signals, automatic identification and classified modulation are performed on underwater signals of active sonar signals and non-cooperative communication signals in real time, and the underwater signals are used for analysis of recovered sea area information. According to the method, deep-level features can be automatically learned from the time-frequency graph, better robustness is achieved for noise and a complex marine environment, and the recognition accuracy is higher.
Need to check novelty before this filing date? Find Prior Art