Method for identifying sound fault based on mel energy spectrum and convolution neural network

A convolutional neural network and fault identification technology, applied in speech analysis, instruments, etc., can solve the problems of large differences in staff discrimination, high detection costs, and slow information transmission speed, achieving strong separability, improving working conditions, The effect of saving manpower

Active Publication Date: 2019-04-09
广州丰石科技有限公司
View PDF6 Cites 30 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0003] The purpose of the present invention is to solve the defects of large discriminative differences among staff members, slow information transmission speed and high detection cost in the prior art, and propose a sound fault identification method based on mel energy spectrum and convolutional neural network

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
  • Method for identifying sound fault based on mel energy spectrum and convolution neural network
  • Method for identifying sound fault based on mel energy spectrum and convolution neural network
  • Method for identifying sound fault based on mel energy spectrum and convolution neural network

Examples

Experimental program
Comparison scheme
Effect test

Embodiment 1

[0043] A sound fault identification method based on mel energy spectrum and convolutional neural network, comprising the following steps:

[0044] S1: Pre-emphasize the voice signal to increase the high-frequency resolution of the sound;

[0045] Pre-emphasizes the speech signal to increase the high-frequency resolution of the sound. The general transfer function of pre-emphasis is H(z)=1-az^-1. The present invention uses a first-order FIR high-pass filter to realize pre-emphasis, wherein a is a pre-emphasis coefficient, and the voice sampling value at n moments is x(n), and the result after pre-emphasis processing is y(n)=x(n) -ax(n-1), take a=0.95 here.

[0046] S2: Framing the voice signal. In terms of timing, a part of the audio data is intercepted at a certain interval to form a frame, and the interval time is the step size of the frame. Since the sound signal has short-term stationary characteristics, framing the audio helps to further subdivide the characteristics of...

Embodiment 2

[0075] A sound fault identification method based on mel energy spectrum and convolutional neural network, comprising the following steps:

[0076] S1: For the input audio data, pre-emphasize y(n)=x(n)-0.95*x(n-1) according to the following formula;

[0077] S2: Take the average value of the two-channel audio and change it to a single channel, and divide the data into frames with a single frame sampling point of 612 and a step size of 306

[0078] S3: Add a window to each frame, the window is a Hamming window, and the coefficient a=0.46

[0079] S4: Perform fast Fourier transform on each frame of data to generate energy spectrum

[0080] S5: passing the energy spectrum through a Mel-scale triangular band-pass filter. The number of filters is 64, the maximum frequency is 22050 (half of the sampling point frequency 44100)

[0081] S6: The data generated by S5, with the frequency domain as the Y axis and the time domain as the X axis, is converted into a Mei energy spectrum

...

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 discloses a method for identifying a sound fault based on a mel energy spectrum and a convolution neural network. The method comprises the following steps: first, performing pre-emphasison initially input audio data; then, performing framing and windowing processing on the data; after that, performing fast Fourier transform on framed and windowed data; extracting energy features ona frequency domain, processing an energy spectrum by using a set of mel-scale triangular filter banks; and after that, converting the data into a Mel energy spectrum by using energy in different frequency domains corresponding to each frame as a Y axis, and different frames in a time domain as an X axis. After that, the energy spectrum needs to be further framed to adapt to an input of a CNN (convolution neural network). Each frame is a sample, and one-hot coding of a label corresponding to each sample is used as an output of the CNN to train a CNN model until a network training error reachesthe minimum. During prediction, a probability value of each type of label is output, and a label with the largest probability value is taken as a final discrimination result.

Description

technical field [0001] The invention relates to the field of AI sound fault detection and recognition, and more specifically relates to a sound fault recognition method based on mel energy spectrum and convolutional neural network. Background technique [0002] Large-scale air-conditioning cooling equipment has been quite popular in practical applications, and manual detection is mostly used for the detection of its failure. The method of patrol inspection by professional staff can rely on the staff's senses and feelings to check the appearance, vibration, operating sound, etc. of the equipment, so as to determine whether the equipment is damaged, loose and other abnormalities. This method is simple and convenient, but the information obtained is very limited, and the knowledge, skills, experience, and observation ability of the staff vary greatly, and the inspection results are not satisfactory. Contents of the invention [0003] The purpose of the present invention is t...

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
Patent Type & Authority Applications(China)
IPC IPC(8): G10L25/51G10L25/30G10L25/21G10L25/18G10L25/15
CPCG10L25/15G10L25/18G10L25/21G10L25/30G10L25/51
Inventor 陈曦蓝志坚陈卓李学辉喻春霞容伯杰
Owner 广州丰石科技有限公司
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

Browse by: Latest US Patents, China's latest patents, Technical Efficacy Thesaurus, Application Domain, Technology Topic.

© 2024 PatSnap. All rights reserved.Legal|Privacy policy|Modern Slavery Act Transparency Statement|Sitemap