Deep-sea direct sound area target depth estimation method based on deep neural network

A deep neural network and target depth technology, which is applied in the field of target depth estimation in the deep-sea direct sound area, can solve the problems of difficulty in giving an accurate solution to the point source interference structure, and estimate the sound source depth error. The effect of limitation, low relative error, strong environmental adaptability

Active Publication Date: 2021-07-13
INST OF ACOUSTICS CHINESE ACAD OF SCI
View PDF4 Cites 1 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

However, it is difficult to give an accurate solution to the point source interference structure in the actual hydrological environment, and the solution results at constant sound velocity may lead to errors in estimating the depth of the sound source

Method used

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
View more

Image

Smart Image Click on the blue labels to locate them in the text.
Viewing Examples
Smart Image
  • Deep-sea direct sound area target depth estimation method based on deep neural network
  • Deep-sea direct sound area target depth estimation method based on deep neural network
  • Deep-sea direct sound area target depth estimation method based on deep neural network

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0066] The technical solutions of the present invention will be described in detail below in conjunction with the accompanying drawings and embodiments.

[0067] The present invention proposes a method for estimating the target depth of deep-sea direct sound area based on deep neural network, which is divided into three main steps: simulation sound field data generation, simulation data preprocessing and marking, simulation data training deep neural network, and measured data preprocessing And the output of the measured data through the neural network.

[0068] Step 1: Calculate the simulated sound pressure field data set in the frequency domain by inputting the environmental parameters into the sound field calculation program KrakenC.

[0069] Step 2: Remove the influence of the sound source amplitude through frequency normalization processing, and then output the beam response data set of the simulation data in the frequency-grazing angle domain through conventional beamform...

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to view more

PUM

No PUM Login to view more

Abstract

The invention provides a deep-sea direct sound area target depth estimation method based on a deep neural network, and the method comprises the steps: carrying out the FFT processing of time-array element domain data obtained through the actual measurement of a vertical array, and obtaining frequency-array element domain data; performing conventional beam forming processing on the frequency-array element domain data after frequency domain normalization processing to obtain a beam response matrix of a frequency-glancing angle domain; and inputting the wave beam response matrix of the frequency-glancing angle domain into a pre-trained deep neural network, and outputting a sound source depth estimation result. The method provided by the invention can overcome the limitation of the multi-spectrum depth estimation method by the frequency band width and the array aperture; and the accuracy of depth estimation is obviously higher than that of a deep neural network estimation method with input data being frequency-depth sound pressure field data and a traditional sound source depth estimation method, and the relative error is the lowest.

Description

technical field [0001] The invention relates to the field of underwater acoustic physics, in particular to a method for estimating the depth of a target in a deep-sea direct sound region based on a deep neural network. Background technique [0002] The depth of the broadband sound source in the deep-sea direct wave region can be estimated by the Multispectral Depth Estimation Method (MSDE Method). This method is based on the point source interference sound field based on the constant sound velocity model, and based on the analysis of the interference spectrum of the array receiving frequency domain signal, Fourier transform, beamforming and improved Fourier transform cubic signal spectrum are respectively performed on the original signal Analysis, there is no need to predict the motion state of the sound source, and the beamforming result also has a certain ability to resist spatial noise interference. However, it is difficult to give an accurate solution to the point sourc...

Claims

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to view more

Application Information

Patent Timeline
no application Login to view more
IPC IPC(8): G01S11/14G06N3/04
CPCG01S11/14G06N3/045
Inventor 王同王文博苏林任群言马力
Owner INST OF ACOUSTICS CHINESE ACAD OF SCI
Who we serve
  • R&D Engineer
  • R&D Manager
  • IP Professional
Why Eureka
  • Industry Leading Data Capabilities
  • Powerful AI technology
  • Patent DNA Extraction
Social media
Try Eureka
PatSnap group products