JPH01282483A - Sound analyser - Google Patents

Abstract

PURPOSE:To automatically analyze sound easily with high accuracy, by excluding the effect due to Doppler effect generated by the relative speed with a moving body to be measured. CONSTITUTION:The data at the time when a relative speed is zero is stored in a standard data memory apparatus 12 as standard data at every moving body (ship, airplane) 13 to be measured. The data obtained when the relative speed with the moving body 13 at the time of measurement is generated is set to data to be measured, and this data and the standard data of the apparatus 12 are respectively converted to high speed Fourier transform by an FFT 16 to calculate the frequency spectrum SP of each data. After the specific SP in the SP of the data to be measured is linearly converted by a linear converter 17, the correlation between the SP of the data to be measured and that of the standard data is taken. When this correlation value is a reference value or less, both of them are regarded to be same to take out the standard data and, when said correlation value is the reference value or less, the next standard data is read to be subjected to Fourier transform and linear conversion along with the data to be measured and correlation is again taken to select standard data wherein the correlation value becomes the reference value or more and the moving body 13 can be discriminated on the basis of this data.

Description

[Detailed Description of the Invention] [Object of the Invention] (Industrial Field of Application) This invention is an acoustic analysis device that individually identifies moving objects (ships, airplanes, etc.) based on the unique sounds emitted by the objects! Regarding.

- (Prior art) Conventional acoustic analysis devices take advantage of the fact that solid bodies emit sounds with individual frequencies, capture the oscillated sound of the solid body, and generate a frequency spectrum (hereinafter simply referred to as spectrum).
Analyze and compare with previously obtained sample data.
If the comparison results in a match, it can be identified as a specific solid. However, in the case of a moving object, there is numerically a relative speed between it and the acoustic analysis device, and this relative speed causes the Doppler effect, which causes the frequency to fluctuate.

When a human conducts acoustic analysis, it is possible to distinguish by canceling this Doppler effect and comparing it with sample data, but this Doppler effect becomes a hindrance when automatic identification is attempted.

In other words, the automatic identification method is performed by ``correlating the spectrum of the measured data with the spectrum of the sample data, and if there is a correlation of more than a certain degree, they are considered to be the same object,'' but the Doppler effect occurs. In this case, the position that gives this maximum correlation is not constant, which poses a problem in subsequent analysis.

(Problems to be Solved by the Invention) As mentioned above, in conventional acoustic analysis devices, when there is a relative velocity between the moving object to be measured (the oscillation source), the Doppler effect occurs, so automatic analysis is significantly difficult. It was difficult.

Therefore, this invention was made to eliminate the above-mentioned drawbacks, and even if the Doppler effect occurs due to the relative velocity with the moving object to be measured, the influence of this Doppler effect is excluded, and the
The purpose is to provide an acoustic analysis device that can perform automatic analysis easily and with high accuracy.

[Configuration of the Invention (Means for Solving the Problems) In order to achieve the above object, the acoustic analysis device according to the present invention has an acoustic signal input that acquires sound waves emitted from a moving object to be measured and converts them into acoustic signals. an analog/digital converter that converts the acoustic signal obtained by the acoustic signal input device into digital data; and output data of the analog/digital converter when the relative velocity of each of the moving objects to be measured is zero. A sample data storage device that stores the data as sample data, and inputs output data of the analog/digital converter when a relative velocity with the moving object to be measured occurs as data to be measured, and also inputs the output data from the sample data storage device as the data to be measured. A fast Fourier transform means for reading out and inputting the sample data, performing fast Fourier transform on the data to be measured and the sample data to obtain a frequency spectrum of each data; a linear conversion means for linearly converting the spectrum of the data to be measured so that a specific spectrum of the spectrum of the data to be measured matches a specific spectrum of the sample data; and a spectrum of the data to be measured and the sample converted by this means. Enter the spectrum of data? & A correlation means that correlates the spectrum of the measured data with the spectrum of the sample data, and when the correlation value obtained by this means is greater than a reference value, the two are regarded as the same and the sample data is taken out, and the correlation of this means is When the value is less than the reference value, the next sample data is read from the sample data storage device, and after performing the fast Fourier transform processing and linear transformation with the data to be measured, the correlation is taken again and the correlation value becomes equal to or greater than the reference value. A sample data selection means for selecting sample data is provided, and the moving body to be measured is identified by the original data selected by the sample data selection means.

(Operation) The acoustic analysis device having the above configuration converts the sound waves emitted from the moving object to be measured into an acoustic signal, and converts this acoustic signal into digital data. Here, the data when the relative velocity is zero for each moving object to be measured is stored as sample data in the sample data storage device. The data to be measured and the sample data read from the sample data storage device are each subjected to fast Fourier transform to obtain the frequency spectrum of each data, and a specific spectrum of the spectrum of the data to be measured matches the specific spectrum of the sample data. After linearly transforming the spectrum of the data to be measured so that the spectrum of the data to be measured and each spectrum of the sample data are correlated, When this correlation value is above the reference value, the two are considered to be the same and the sample data is taken out. When the correlation value of this means is below the reference value, the next sample data is read out from the sample data storage device and together with the data to be measured. After performing the fast Fourier transform processing and linear transformation, correlation is again taken and sample data whose correlation value is equal to or greater than the reference value is selected. The moving object to be measured can be identified using the sample data selected in this way.

(Example) An embodiment of the present invention will be described below with reference to the drawings.

FIG. 1 shows its configuration, and 11 is a host computer X81 that collectively controls each component of this device. This may simply be a controller. Reference numeral 12 is a sample data storage device in which sample data of the moving body to be measured in a stationary state is stored in advance.

Reference numeral 13 denotes a moving object to be measured (an oscillation source), and the sound waves emitted from this moving object 13 are converted into electrical signals by a microphone 14, and digitized by an analog/digital (A/D>converter 15). A filter (not shown in FIG. 1) is provided in front of the A/D converter 15 to select only a necessary frequency band.

The A/D converted signal is FFTIG as the data to be measured.
sent to.

This FFT16 converts the measured data into a fast Fourier transform (F
FT), and inputs the sample data read out from the sample data storage device [12] through the host machine 11.
This sample data is subjected to fast Fourier transform, and the resulting spectra of the measured data and the sample data are both sent to the linear converter 17.

This linear converter 17 linearly transforms the spectrum of the input data to be measured so that a specific spectrum (for example, the lowest frequency with a level higher than a certain room) of the spectrum of the input data to be measured matches the specific spectrum of the sample data. The spectrum of the measured data converted here is sent to the correlator 18 along with the spectrum of the sample data.
sent to.

The correlator 18 correlates the spectrum of the data to be measured and the spectrum of the sample data.

If the correlation value is greater than the reference value, the two are considered to be the same,
When the value is below the reference value, the next sample data is read from the sample data storage device 12 and subjected to the above FFT together with the measured data.
After processing and linear transformation, correlation is again taken and sample data whose correlation value is greater than or equal to the reference value are selected.0 The selected sample data are sent to the display 19. This display 19
Displays the identification result of the moving object to be measured from the input sample data.

The operation of the above configuration will be described below with reference to FIG. Although it is actually digital processing, the second
The figures are shown in analog form for ease of explanation.

Now, both the moving object 13 to be measured and the acoustic analysis device are in a stationary state, and the acoustic signal emitted from the moving object 13 to be measured is captured by the microphone 14 to obtain a frequency signal as shown in FIG. 2(a). Suppose that FFT this frequency signal
16, a spectrum as shown in FIG. 16(b) is obtained.

The data obtained by the A/D converter 15 at this time is stored in advance in the sample data storage device 12 as sample data.

Next, when the moving object 13 to be measured moves, a Doppler effect occurs due to the relative speed, and the frequency of the signal captured by the microphone 14 fluctuates. When this frequency signal is used as data to be measured and is subjected to fast Fourier transform using FFT 16, a spectrum as shown in FIG. 2(c) is obtained. In other words, as can be seen by comparing it with the spectrum of the sample data in FIG. 2(b), the spectrum of the measured data is shifted from the spectrum of the sample data by the Doppler effect. Conventionally, due to this shift, the comparison between the two is a fixed number,
The position that gave the maximum correlation was not constant, which caused problems in automatic analysis.

Therefore, in this invention, the spectra of the sample data and the data to be measured that have been fast Fourier transformed by the FFT 16 are input to the linear converter 17, and as shown in FIG. The spectrum of the measured data is linearly transformed so that the lowest frequency Rf1 with a level shift level matches the specific spectrum of the sample data (in the figure, the lowest frequency f01 with a constant level shift level).

This linear conversion scans the spectrum of the measured data in a fixed frequency band to match the specific spectrum of the sample data, removes the shift due to the Doppler effect, and converts the data into data when the moving object is stationary. To explain specifically, first, the frequency f of the observed data to be measured and the frequency f of the sample data.

Using the relationship of f = fo ・A (A is a constant), find the constant A by the above search, and find this constant A.
This is realized by calculating f/A=fO using , and converting the measured data to the frequency at rest.

If the specific spectrum of the measured data obtained by the linear transformation matches the specific spectrum of the sample data, it means that the shift due to the Doppler effect has been removed. Therefore, the spectrum of the measured data after linear conversion and the spectrum of the sample data are sent to the correlator 18 and correlated. If a correlation above the standard value is obtained, it is assumed that they are the same and the mobile object identification name is determined from the sample data.
It is displayed on the display 19. If the value is less than the reference value, the original data is changed, the value of the constant A is determined again, the correlation is taken by linear transformation, and this operation is automatically repeated until a correlator having the value equal to or greater than the reference value is obtained.

Therefore, since the acoustic analysis device having the above configuration replaces the data to be measured while the moving object is moving with the data when the moving object is stationary, it is possible to eliminate the influence of the Doppler effect.
A moving object can be identified easily and with high accuracy by automatically taking a correlation.

In the above embodiment, dedicated hardware is used for each individual process, but it goes without saying that all processes can be performed by the host computer 11 after A/D conversion.

[Effects of the Invention] As described above, according to the present invention, even if the Doppler effect occurs due to the relative velocity with the moving object to be measured, the influence of the Doppler effect can be excluded and automatic analysis can be performed easily and with high precision. An acoustic analysis device can be provided.

[Brief explanation of the drawing]

FIG. 1 is a block circuit diagram showing an embodiment of an acoustic analysis device according to the present invention, and FIG. 2 is a diagram for explaining the operation of the embodiment. DESCRIPTION OF SYMBOLS 11... Host computer, 12... Sample data storage device, 13... Measured moving object, 14... Microphone, 15... A/D converter, 16... FFT, 17...
... Linear converter, 18 ... Correlator, 19 ... Display device. Applicant's agent Patent attorney Takehiko Suzue

Claims (1)

[Claims] an acoustic signal input device that acquires a sound wave emitted from a moving object to be measured and converts it into an acoustic signal; an analog/digital converter that converts the acoustic signal obtained by the acoustic signal input device into digital data; and the device to be measured. a sample data storage device that stores output data of the analog/digital converter when the relative velocity of each moving object is zero as original data;
The output data of the analog/digital converter when a relative velocity with the moving body to be measured occurs is input as the data to be measured, and the sample data is read from the sample data storage device and inputted, and the data to be measured and the sample are inputted. A fast Fourier transform means that performs fast Fourier transform on each data to obtain the frequency spectrum of each data, and each spectrum of the data to be measured and sample data obtained by this means is input, and a specific spectrum of the spectrum of the data to be measured is input. A linear transformation means linearly transforms the spectrum of the data under test so that it matches a specific spectrum of the sample data, and a spectrum of the data under test converted by this means and a spectrum of the sample data are input, and the spectrum of the data under test is converted into the spectrum of the data under test. Correlation means that calculates the correlation with the spectrum of the sample data, and when the correlation value obtained by this means is equal to or higher than the reference value, the two are considered to be the same and the sample data is extracted. When the next sample data is read from the sample data storage device, and after performing the fast Fourier transform processing and linear transformation together with the data to be measured, the sample data is correlated again and the sample data whose correlation value is equal to or greater than the reference value is selected. 1. An acoustic analysis apparatus, comprising: a selecting means, and identifying the moving body to be measured based on the sample data selected by the sample data selecting means.