JP3867577B2 - Vibration wave determination device and vibration wave determination method - Google Patents

Description

[0001]
BACKGROUND OF THE INVENTION
The present invention relates to a vibration wave determination device and a vibration wave determination method for determining the state of an object based on vibration waves generated by the object.
[0002]
[Prior art]
Conventionally, there is a method of tapping a target object with a hammer or the like, detecting a sound or vibration generated at that time, and determining the content of the target object, inspection of a crack of the structure, etc. based on the vibration wave. It has been implemented. In this specification, sound and vibration are collectively referred to as vibration waves.
[0003]
A method for automatically carrying out this method is described in JP-A-10-300730.
In the present invention, a microphone and a vibration detector are attached to a hammer that strikes a product such as a can. The sound detected by the microphone is separated and extracted for each frequency band by an analog bandpass filter. On the other hand, a trigger signal is acquired by the vibration detected by the vibration detector. A correction value is acquired based on the amplitude value of the vibration.
[0004]
The signal level of each divided frequency band is held by the peak hold circuit triggered by the trigger signal and input to the corrector. The signal for each frequency band input to the corrector is corrected by the correction value. The corrected signal is compared with a reference signal by a comparator. When the comparison signal value for each frequency band deviates from the predetermined relationship, the determiner determines that there is an abnormality in the product and sends a signal to the alarm device.
[0005]
[Problems to be solved by the invention]
In the above conventional method, the determination is made using a preset internal reference value. For this reason, when the frequency or sound pressure of the sound varies depending on the product, it is necessary to change the reference value every time the type of the product is changed.
Further, since all frequency bands prepared in advance are analyzed, there is a possibility that unnecessary calculation may occur even if it is sufficient to make a determination only for a specific frequency band.
[0006]
An object of the present invention is to make it possible to flexibly cope with a change in an object in a vibration wave determination apparatus and a vibration wave determination method for determining the state of an object using vibration waves.
[0007]
[Means for Solving the Problems]
The present invention has been made to achieve the above object. When the object is assembled, different vibration waves are generated depending on whether the object is normally assembled or abnormally assembled. Also, different vibration waves are generated depending on the presence / absence of contents and the presence / absence of scratches. In the vibration wave determination apparatus of the present invention, when a vibration wave generated from an object is input, the frequency separation means extracts a vibration wave in the frequency band stored in the parameter table, and the coefficient stored in the parameter table. It is corrected by. The determination means determines the defect of the object based on the corrected signal level. The vibration wave determination method of the present invention detects a vibration wave of an object, extracts vibration waves of a plurality of frequency bands from the detected vibration wave, and determines the signal levels of the extracted plurality of frequency bands as the signal. Correction is made based on the coefficient corresponding to the level, and the vibration wave is determined based on the sum of the corrected signal levels.
[0008]
The parameter table stores a frequency band used for determination and a coefficient corresponding to the frequency band for each object. Generally, the frequency band to be confirmed differs depending on the object. For this reason, by storing the pass / fail judgment condition for each object in the parameter table, the system can be simplified and the operation speed can be increased.
[0009]
The present invention can be provided with trigger means for commanding the start of operation. As a result, the vibration wave determination device does not start the determination operation until a command is issued, so that there is no possibility that ambient noise when there is no command is erroneously processed as a detection signal. Therefore, the influence by these can be prevented and the reliability of comparison is improved.
[0010]
According to the present invention, it is possible to prevent the outflow of defective products and the like by determining the state of a portion in a part that cannot be seen by being hidden by the vibration wave generated by the object.
Then, by using the parameter table, it is possible to easily select a frequency band suitable for determination of an object. Therefore, even when the type of the object changes, it is possible to respond flexibly.
Furthermore, since operations for unnecessary frequency bands that are not in the parameter table are omitted, high-speed processing is possible.
[0011]
DETAILED DESCRIPTION OF THE INVENTION
Embodiments of the present invention will be described with reference to the drawings.
In the following description, an example in which a product defect is determined by detecting a sound generated during product assembly work will be described. However, the vibration wave determination apparatus of the present invention is not limited to this application, and can be applied to other applications. For example, it can be used for sound inspection such as discrimination of the contents of a container and inspection of scratches on structures. Further, the present invention can be applied to vibration detection instead of sound detection.
[0012]
(Embodiment 1)
FIG. 1 shows a system configuration in the first embodiment of the present invention.
A signal from the microphone 2 is input to the vibration wave determination device 1. The microphone 2 detects the vibration generated in the product 3 that is the object of vibration wave determination as a sound wave and converts it into an electrical signal. The electric signal of sound pressure input from the microphone 2 is amplified by the amplifier 4 and output to the A / D converter 5.
[0013]
When receiving the trigger signal from the trigger detector 6, the A / D converter 5 starts processing and converts the sound pressure signal into a digital signal. The trigger detector 6 is command means for informing the start of the assembly operation, and is provided so as not to erroneously detect noise when there is no command. A specific example of the trigger detector 6 is, for example, an OFF signal at the return side end in the case of cylinder assembly, and a push button operated by an operator in the case of manual assembly.
[0014]
The digital signal output from the A / D converter 5 is input to a wavelet transform operator 7. The wavelet transform computing unit 7 separates the digital signal for each frequency band specified by the parameter table 9 and converts it into a time series signal.
The wavelet transform computing unit 7 is a computing unit that separates the digital sound pressure signal into time-series signals of each frequency by enlarging or reducing the basis function (wavelet function). In this example, the frequency separation characteristics and the reliability of determination are improved by selecting an appropriate basis function according to the instruction of the parameter table 9 in accordance with the type of product during this calculation.
[0015]
In accordance with the product number input from the product number information input means 10, the parameter table 9 outputs a frequency band effective for determining the product number to the wavelet transform computing unit 7 and outputs coefficients to the correctors 81, 82.
FIG. 2 shows the contents of the parameter table 9. The parameter table 9 stores a frequency band and a gain (correction coefficient) used for determination for each product (part numbers A to C). A method for creating the parameter table will be described later.
[0016]
The parameter table will be described with reference to FIGS.
FIG. 3 shows an operation process in which a product 3 having a certain product number is assembled by snap fitting. (A) shows a state before two parts are assembled, (B) shows a state where the assembly has been completed normally, and (C) shows a state where the two parts are not completely assembled. A different sound is generated at the time of assembly between the case of moving from (A) to (B) and the case of moving from (A) to (C).
[0017]
The sound pressure waveform when the assembly is normally performed and the assembly sound is detected is shown in the uppermost part of FIG. 4A, and the assembly operation is not performed normally in the uppermost part of FIG. The sound pressure waveform when no assembly sound is detected is shown. In addition, the waveform displayed as the mechanical sound of the equipment in the figure indicates that the microphone 2 detects a sound irrelevant to the assembling work.
[0018]
The lower part of FIG. 4 shows a sound pressure waveform obtained by separating the waveform shown in the uppermost part for each predetermined frequency band.
When (A) and (B) are contrasted, it is recognized that there are frequency bands whose waveforms vary greatly depending on the presence or absence of the assembly sound. One or a plurality of such frequency bands are selected and stored in the parameter table 9. At this time, a maximum value is obtained for each frequency band, and a correction coefficient necessary for weighting is calculated.
[0019]
Returning to FIG. 1, the wavelet transform computing unit 7 performs frequency separation for the frequency bands stored in the parameter table 9, and the resulting time series signals are input to different correctors 81, 82.
The correctors 81, 82,... Weight each frequency band using the gain (correction coefficient) stored in the parameter table 9. When the wavelet transform calculator 7 performs weighting, the correctors 81, 82,... Can be omitted.
[0020]
FIG. 5 shows the output signal (A) of the wavelet transform computing unit 7 and the output signals (B) of the correctors 81, 82. In the illustrated example, it is shown that the wavelet transform computing unit 7 performs frequency separation for three frequency bands. In addition, it is shown that the signal levels are almost equalized by weighting by the correctors 81, 82.
[0021]
The corrected sound pressure signal is input to the determiner 11, and the quality of the product 3 is determined (determining whether there is an assembly sound). As an example of the judgment method, the sum of signal levels for each frequency band is obtained and compared with a predetermined threshold value. If the threshold value is exceeded, it is normal (with built-in sound), and if it is below the threshold value, it is abnormal. It is determined that there is no assembly sound.
Based on the determination result, the determination device 11 displays the result on the display device 12, and outputs the result to the alarm device 13 in the case of failure.
[0022]
According to the example described above, even if the frequency and sound pressure of the sound fluctuate depending on the type of the product 3, it can be easily handled by using the frequency band and coefficient for each product stored in the parameter table 9 in advance. can do. Further, unnecessary processing such as signal separation and correction is not performed for frequency bands not stored in the parameter table 9. Therefore, it is possible to realize high speed computation.
[0023]
Further, by removing minute vibrations in the waveform separated for each band, it is possible to accurately determine by removing noise components.
When vibration is detected instead of sound, a vibration sensor is brought into direct contact with the product 3 to convert the vibration of the product 3 into an electric signal and input it to the amplifier 4.
[0024]
(Embodiment 2)
In this example, an image processing method is used in combination as a confirmation method in the assembly process. According to this example, it is possible to realize a variety of products, including products that do not generate assembly sound (or have low sound), without missing.
The principle of determining the assembly part from an image will be described.
As shown in FIG. 3 described above, when the assembly is normally completed from the state before the assembly shown in (A), the gap amount between the members becomes 0 as shown in (B). Further, if the assembly is not sufficiently performed, a gap A is generated between the members as shown in (C).
[0025]
FIG. 6 shows a system configuration in the second embodiment of the present invention. The configuration in FIG. 6 has many parts in common with the system configuration in FIG. 1 described above. Therefore, in the following description, the description overlapping with the description in FIG. 1 is omitted, and the configuration unique to FIG. 6 will be mainly described.
The configuration for performing determination processing using sound, that is, the configuration from the microphone 2 to the determination unit 11 is the same as that in FIG. In the determination unit 11, the sound determination is performed by the sound determination unit 14.
[0026]
The state of the product 3 is photographed by the camera 21 for image processing. An image photographed by the camera 21 is converted into an electric signal and taken into the image input circuit 22. When the assembly is completed, a trigger signal is input from the trigger detector 6 to the image input circuit 22, and an image at that time is output to the image determination unit 23 of the determiner 11.
The image determination unit 23 extracts the gap length by performing image processing on the captured image information.
[0027]
A threshold value for pass / fail judgment is acquired from the parameter table 9.
FIG. 7 shows the contents of the parameter table 9 used in this example.
The threshold value used in this example is stored in the product number B as the gap length L1. In the product numbers A and C, a frequency band and a gain for determination by sound are stored.
[0028]
By specifying the product number B with the product number information input means 10, the gap length L 1 is output from the parameter table 9 to the determiner 11. The image determination unit 23 compares the gap length extracted from the image with the gap length L1 obtained from the parameter table 9, and determines that the image is normal if the image gap length is equal to or smaller than the threshold value. Judged to be abnormal.
[0029]
Based on the determination result, the determination device 11 displays the result on the display device 12, and outputs the result to the alarm device 13 in the case of failure.
According to this example, the sound quality waveform is used for sound generation, and the quality of the assembly process is determined using an image for sound generation. As a result, the determination can be made more reliable.
[0030]
Instead of detecting the gap length by image processing, the determination can also be made by detecting the position (or position variation) of the member. In this case, a detector can be brought into contact with a predetermined place of the product 3, and the assembly depth and the nail lift can be detected based on the position information obtained from the detector.
[0031]
(Embodiment 3)
This example implements a method for logging the determined data. By logging the judgment data for each product, it can be used for analyzing the mechanism of defect occurrence and preventing the recurrence of defects.
FIG. 8 shows a system configuration. The system configuration of this example includes all the system configurations of FIG. 6 described above, and further adds new components. However, in FIG. 8, a part of the configuration shown in FIG. 6 is omitted. Moreover, in the following description, the description which overlaps with the description of FIG. 6 is abbreviate | omitted, and demonstrates a structure peculiar to FIG.
[0032]
The determiner 11 determines the quality of the product based on sound or an image, and outputs the result to the logging data generation unit 31. The logging data generation unit 31 includes a sound data processing unit 32 and an image data processing unit 33, to which determination results from the sound determination unit 14 and the image determination unit 23 are input, respectively. The sound data processing unit 32 and the image data processing unit 33 further take in sound data and image data from the A / D converter 5 and the image input circuit 22.
[0033]
The logging data generator 31 processes the sound data and the image data, adds the determination result obtained from the determiner 11, the time data obtained from the trigger detector 6, and the product number data obtained from the parameter table 9, Output to the logging device 34. The logging device 34 stores the data from the logging data generation unit 31 in the logging output storage unit 35.
[0034]
(Embodiment 4)
A configuration for generating the parameter table 9 will be described with reference to FIG.
The configuration of the vibration wave determination apparatus 1 in FIG. 9 is the same as that of the system in FIG. In this example, a parameter table generator 41 is connected to the output side of the determiner 11. Further, an HMI device 42 is connected to the output side. The HMI device 42 adds the result data to the parameter table 9.
In the following description, the description overlapping with the description of FIG. 6 is omitted, and the configuration unique to FIG. 9 will be mainly described.
[0035]
A case where determination is performed based on sound will be described.
First, the assembly work of the product 3 is performed, a sound with an assembly sound is generated a plurality of times, and the assembly sound is detected by the microphone 2.
When generating the parameter table, the wavelet transform computing unit 7 performs frequency separation for all the frequency bands shown in FIG.
[0036]
Each time an assembly sound is generated, frequency separation is performed by the wavelet transform computing unit 7. This data passes through the determiner 11 and is input to the parameter table generator 41. The parameter table generator 41 accumulates data for a plurality of times, and calculates the average and variance of the maximum values for each frequency band.
Next, a plurality of sound data are input in a state where no assembly sound is generated in the assembly operation, and the average and variance of the maximum values are similarly obtained for each frequency band.
[0037]
The parameter table generator 41 extracts, for each frequency band, a frequency band in which data is greatly different depending on whether or not an assembly sound is present, and selects this as a determination frequency band. A weighting amount (coefficient) is calculated based on the data level of the selected frequency band. The resulting frequency band and coefficient are output to an HMI (Human Machine Interface) device 42.
[0038]
When the operator determines a threshold value from the indicated data in the HMI device 42 and adds product number information, the operator instructs registration processing in the parameter table 7. Thereby, data about a new product number can be efficiently created in the parameter table 7.
The product number information can be input by using a signal from a control device such as a PLC instead of a direct operation from an operator.
When inputting an image feature quantity to the parameter table 7, the operator inputs data such as a product number and a feature quantity directly from the HMI device 42.
[0039]
【The invention's effect】
ADVANTAGE OF THE INVENTION According to this invention, in the vibration wave determination apparatus which determines the state of a target object with a vibration wave, it can respond flexibly to the change of a target object.
[Brief description of the drawings]
FIG. 1 is a diagram showing a system configuration according to a first embodiment of the present invention.
FIG. 2 is a diagram showing the contents of a parameter table in FIG.
FIG. 3 is a diagram showing an assembly process of the product of FIG. 1;
4 is a diagram showing a sound pressure waveform detected in the process of FIG. 3;
5 is a diagram showing a state of conversion of the sound pressure waveform in FIG. 1. FIG.
6 is a diagram showing a system configuration of Embodiment 2. FIG.
7 is a view showing the contents of the parameter table of FIG. 6;
FIG. 8 is a diagram showing a system configuration of Embodiment 3 of the present invention.
FIG. 9 is a diagram showing a system configuration of Embodiment 4 of the present invention.
[Explanation of symbols]
DESCRIPTION OF SYMBOLS 1 ... Vibration wave determination apparatus 2 ... Microphone 3 ... Product 4 ... Amplifier 5 ... A / D converter 6 ... Trigger detector 7 ... Wavelet transformation calculator 81-82 ... Corrector 9 ... Parameter table 10 ... Product number information input means 11 Determinator 12 ... Display 13 ... Alarm 14 ... Sound determination unit 21 ... Camera 22 ... Image input circuit 23 ... Image determination unit 31 ... Logging data generation unit 32 ... Sound data processing unit 33 ... Image data processing unit 34 ... Logging Device 35 ... Logging data storage unit 41 ... Parameter table generator 42 ... HMI device

Claims (8)

A vibration wave input means to which a signal from a vibration wave detection means for detecting a vibration wave of the object is input;
A parameter table storing a plurality of frequency bands and coefficients corresponding to the frequency bands;
Frequency separation means for extracting vibration waves in a plurality of frequency bands indicated by the parameter table from the input vibration waves;
Correction means for correcting the signal levels of the plurality of separated frequency bands based on the coefficients stored in the parameter table;
Determining means for determining a vibration wave based on a sum of signal levels of the corrected plurality of frequency bands ;
A vibration wave determination apparatus comprising: Comprising trigger input means for inputting a signal from the trigger means;
The vibration wave determination device according to claim 1, wherein the frequency separation unit is a wavelet calculation unit, and starts the separation of the frequency band when the trigger signal is input. Comprising an image input means for inputting an image signal from an imaging means for taking an image of an object;
The vibration wave determination apparatus according to claim 1, wherein the determination unit extracts a feature amount from the input image signal and determines a target based on the feature amount.   The vibration wave determination apparatus according to claim 1, further comprising a storage unit that stores determination data output from the determination unit. A parameter table generator connected to the output side of the determination means;
The parameter table generator creates a frequency band and the coefficient for the object based on a comparison between determination data when the vibration wave from the object is normal and determination data when the wave is abnormal. The vibration wave determination apparatus according to claim 1, wherein the vibration wave determination apparatus is stored in the parameter table. The parameter table stores, for each object, a frequency band used for determination and a coefficient corresponding to the frequency band,
6. The vibration wave determination device according to claim 1, wherein the parameter table selects a frequency band suitable for determination of an object in accordance with a product number input from a product number information input unit. . Detect the vibration wave of the object,
Extracting vibration waves in a plurality of frequency bands from the detected vibration waves,
Correcting the extracted signal levels of the plurality of frequency bands based on a coefficient corresponding to the signal level;
A vibration wave determination method for determining a vibration wave based on the sum of the corrected signal levels. 8. A frequency band suitable for determination of an object is selected in accordance with a product number input from a product number information input unit, and vibration waves in a plurality of frequency bands are extracted from the detected vibration wave. The vibration wave determination method described.

Images