Open AccessTime-Frequency Feature-Based Underwater Target Detection with Deep Neural Network in Shallow Sea
Author(s) -
Yunliang Zheng,
Qiyong Gong,
Shufang Zhang
Publication year - 2021
Publication title -
journal of physics. conference series
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.21
H-Index - 85
eISSN - 1742-6596
pISSN - 1742-6588
DOI - 10.1088/1742-6596/1756/1/012006
Subject(s) - computer science , signal (programming language) , feature (linguistics) , artificial neural network , artificial intelligence , time–frequency analysis , underwater , pattern recognition (psychology) , detector , detection theory , energy (signal processing) , set (abstract data type) , deep learning , data set , computer vision , mathematics , telecommunications , geology , statistics , philosophy , linguistics , oceanography , filter (signal processing) , programming language
In this paper, we propose a time-frequency feature-based signal detection method with deep neural network and apply it for target detection in shallow sea. The short time Fourier transform is employed to capture the time-frequency feature of the target signal. We input the time-frequency diagram of the signal into the deep neural network model, train the network model through the labeled data samples, and use the test set to estimate the signal detection probability. In addition, we analyze the influence of different learning rates on the detection performance. The experimental results show that the performance of the proposed method is better than that of the energy detector, and the detection performance of different learning rates is different.