Open AccessAn estimation method of underwater target spectrum line parameters based on convolutional neural network
Author(s) -
Zhaorui Luo,
Sidan Du,
Yuechao Chen
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 - 10.1088/1742-6596/2010/1/012078
Subject(s) - underwater , convolutional neural network , convolution (computer science) , noise (video) , computer science , sensitivity (control systems) , line (geometry) , artificial neural network , artificial intelligence , spectrum (functional analysis) , pattern recognition (psychology) , algorithm , basis (linear algebra) , acoustics , mathematics , electronic engineering , physics , engineering , oceanography , geometry , quantum mechanics , image (mathematics) , geology
The spectrum line parameters of underwater target radiated noise are important basis for underwater target recognition. In this paper, the convolution neural network (CNN) is applied to estimate the spectrum line parameters. The low frequency analysis record (LOFAR) spectrum of radiated noise is used as input data, and a suitable convolution neural network is constructed to estimate the spectrum line parameters. The processing results of simulation data show that the estimation sensitivity of convolution neural network reaches 91.6% under the condition of low SNR. It proves the effectiveness of CNN in estimating spectrum line parameters of underwater targets.