JP2004085455A - Method and survey instrument for searching sound source of abnormal sound - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To easily and precisely specify a sound source of an abnormal sound. <P>SOLUTION: A sound waveform collecting part 22 of a processing part 14 collects a sound waveform in the present when an equipment is operated, from three or more of microphones 12 arranged in different positions within a sound source searching area A. A comparison part 26 compares the present waveform with a reference sound waveform stored preliminarily in a reference sound waveform storage part 24 to extract only an abnormal sound waveform, and a residual information extracting part 28 removes only amplitude information from the abnormal sound waveform to extract residual information wherein a noise by attenuation of sound pressure generated in propagation of the abnormal sound is removed. A time difference calculating part 30 gets a mutual relation between the residual informations collected by the two microphones 12 to find a time difference up to arrival of the abnormal sound from the same sound source to the respective microphones 12, and a sound source specifying means 32 specifies a sound source position of the abnormal sound, based on the found time difference. <P>COPYRIGHT: (C)2004,JPO

Description

[0001]
TECHNICAL FIELD OF THE INVENTION
The present invention relates to an abnormal sound source searching method and a sound source searching apparatus, and more particularly, to an abnormal sound source searching method and a sound source searching method capable of accurately specifying a sound source of an abnormal sound occurring in a predetermined sound source searching area. It relates to improvement of the device.
[0002]
[Prior art]
It is necessary to quickly detect abnormalities in the operating state of equipment including equipment and devices in factories and the like. Normally, when an abnormality of a device is detected, an operation abnormality is detected by a self-diagnosis function of the facility or device, and an alarm output process, an operation stop process, and the like are performed. Also, in equipment that includes a mechanical drive unit, friction increases due to wear, shape deformation, lack of lubricant, etc. in the motor and bearing parts that constitute the drive unit before actual operation failure occurs. However, abnormal sounds that do not occur during normal operation may be generated. Such an abnormal sound does not affect the normal operation of the device at an early stage, but the influence gradually increases, causing a sudden alarm output or an operation stop.
[0003]
In many cases, such abnormal sound is detected by a site manager or the like recognizing the abnormal sound by his / her own hearing, and further searching for an abnormal sound generating position. In some cases, the abnormal sound is detected by a detector. The method using this detector collects, for example, the current sound waveform currently being emitted in a facility where the equipment is installed by using a plurality of microphones, and collects the collected current sound waveform and an abnormal sound that has been collected in advance. By performing a comparison with a reference sound waveform that does not include the abnormal sound, the occurrence of an abnormal sound is detected, and the time difference between the abnormal sound source reaching the microphone and the abnormal sound is determined. Is calculated backwards. Such an abnormal sound detection system is disclosed in, for example, JP-A-2000-214052. According to this system, it is possible to search for an abnormal sound source without being affected by surrounding sounds.
[0004]
[Problems to be solved by the invention]
However, since the distances from each microphone to the abnormal sound source are different from each other, the waveform reaching each microphone is attenuated, so that different waveforms (different sounds) are generated despite the waveforms emitted from the same sound source. It may be mistakenly recognized as being. In other words, if the comparison of a plurality of abnormal sound waveforms is to be performed as it is, the attenuation generated during the propagation of the sound becomes a noise component, and the correlation between the two waveforms to be compared despite the waveform of the sound emitted from the same sound source. This is because the property is reduced. Furthermore, when a plurality of abnormal sound sources exist in the same sound source search area, a plurality of waveforms recognized as abnormal sounds exist, and the reliability of correlation accuracy for each waveform is further reduced. When there is one sound source generating an abnormal sound, the time difference can be calculated assuming that the sound is an abnormal sound from the same sound source even if the correlation is low. However, this processing becomes a new error factor of the time difference. . In addition, when there are many sound sources emitting abnormal sounds, there is a problem that the reliability of the correlation is further reduced and an accurate time difference cannot be calculated.
[0005]
As a result, it is not possible to accurately calculate the time difference from when the abnormal sound arrives at each microphone from the sound source, and as a result, there is a problem that the reliability of the identification accuracy of the abnormal sound source is reduced.
[0006]
The present invention has been made in view of the above problems, and an object of the present invention is to provide a method and an apparatus for searching for a sound source of an abnormal sound that can easily and accurately specify the sound source of the abnormal sound.
[0007]
[Means for Solving the Problems]
In order to achieve the above-mentioned object, the present invention is generated when a device is operated in a sound source search area using three or more microphones arranged at different positions in the sound source search area. Collecting a current sound waveform, and extracting an abnormal sound waveform by determining the difference between the reference sound waveform collected by each microphone and the current sound waveform when the device is operating normally in the sound source search area, Excluding only the amplitude information of the extracted abnormal sound waveform to obtain residual information that does not include the amplitude information, and cross-correlating the residual information based on the current sound waveform collected by any two microphones, and Determining the time difference before the abnormal sound reaches each microphone, and specifying the sound source position of the abnormal sound based on the time difference obtained by combining the microphones Characterized in that it comprises a step, a.
[0008]
Further, in order to achieve the above object, the present invention collects a current sound waveform that is arranged at a different position in a sound source search area and is generated when a device is operating in the sound source search area. At least three microphones, a reference sound storage unit that stores a reference sound waveform generated when the device is operating normally in the sound source search area, and a current sound waveform collected at the same position for each microphone. A comparison unit that compares with a reference sound waveform to extract an abnormal sound waveform, a waveform processing unit that eliminates only amplitude information from each extracted abnormal sound waveform to obtain residual information, and a waveform processing unit that collects information using any two microphones A time difference calculation unit that obtains a cross-correlation of residual information based on a sound waveform to calculate a time difference at which an abnormal sound from the same sound source reaches each microphone, and a time difference calculation unit that calculates a time difference based on a combination of each microphone. There are, characterized in that it comprises a sound source specifying section for specifying a sound source position of the abnormal sound, the.
[0009]
Here, the current sound waveform is a waveform composed of all sounds present in the sound source search area, and is a normal sound that is emitted when the device is operating normally and an abnormal sound that is emitted when the device is operating abnormally (leading to abnormal operation). (Including the sound of the preceding sign stage) and the background sound existing in the sound source search area. On the other hand, the reference sound waveform is a waveform composed of all sounds existing in the sound source search area, including background sounds and normal sounds existing in the sound source search area, but not including abnormal sounds. Not something. This reference sound is preferably obtained, for example, when it is confirmed that the equipment operates normally, for example, after the simultaneous maintenance of all the equipment is completed. The devices existing in the sound source search area include all equipment and devices that generate sound existing in the area.
[0010]
According to the above configuration, since cross-correlation is performed using residual information excluding amplitude information that causes noise, recognition accuracy of mutual information correlation is improved, and abnormal sounds from the same sound source reach each microphone. Time difference can be obtained with high accuracy. As a result, the position of the sound source of the abnormal sound can be accurately specified. In this case, even when the number of abnormal sound sources is one or more than one, the correlation between individual abnormal sounds can be accurately grasped, so that an accurate sound source search can be performed.
[0011]
Further, in order to achieve the above object, the present invention provides the above configuration, in which, based on the specified sound source position, a step of specifying a corresponding position of a device present in the sound source search area; And reading and presenting information on the device from the storage unit.
[0012]
Further, in order to achieve the object as described above, the present invention, in the above configuration, a search unit for searching for information on the corresponding position of the device in the sound source search area corresponding to the specified sound source position, and the searched information And a presenting unit for presenting.
[0013]
Here, the information on the corresponding position of the device includes, for example, information of the device itself, details of the component of the device (part model, drawing, specifications, etc.), a history of the last replacement time of the component, life information, and the like. It includes the next scheduled replacement time.
[0014]
According to this configuration, the sound source of the abnormal sound can be clearly presented to the administrator, and quick maintenance can be performed.
[0015]
Further, in order to achieve the above object, the present invention, in the above configuration, based on the specified sound source position, controlling the imaging means for obtaining an image of the corresponding position of the sound source in the sound source search area, Presenting the acquired image.
[0016]
In addition, in order to achieve the above object, the present invention provides, in the above configuration, an imaging unit that acquires an image of a sound source corresponding position in a sound source search area based on the specified sound source position, and presents the captured image. And a presentation unit that performs the operation.
[0017]
According to this configuration, the sound source of the abnormal sound can be more clearly presented to the administrator, and efficient maintenance can be performed.
[0018]
BEST MODE FOR CARRYING OUT THE INVENTION
Hereinafter, preferred embodiments of the present invention (hereinafter, referred to as embodiments) will be described with reference to the drawings.
[0019]
FIG. 1 is a configuration block diagram illustrating a configuration concept of a sound source searching device 10 that realizes the abnormal sound source searching method of the present embodiment. The sound source search device 10 includes at least three microphones 12 (in FIG. 1, three microphones 12a, 12b, and 12c are used as an example) to collect sound existing in the sound source search region A at an arbitrary position in the sound source search region A. And a processing unit 14 for specifying a sound source of an abnormal sound emitted by a device such as a device or facility existing in the sound source search area A based on the collected current sound waveform, and presentation of the sound source of the specified abnormal sound. And a display unit 16 as a presentation unit for presenting information about the sound source. The sound source search area A means, for example, a specific range of a process room or a machine room where a plurality of drive-type facilities and devices such as machine tools and assembly robots are arranged in a factory, and the microphone 12 is a facility or the like. The sound (current sound waveform) in the sound source search area A when the is operating is collected. Here, the current sound waveform is a waveform composed of all sounds existing in the sound source search area A, and is a normal sound that is emitted when the equipment or the like is operating normally and an abnormal sound that is emitted when the equipment or the like is operating abnormally (abnormal operation). And the background sound existing in the sound source search area A is also included.
[0020]
In addition, the sound source detection device 10 of the present embodiment uses image information of a position serving as a sound source of an abnormal sound, for example, a video of equipment serving as a sound source, and more detailed components such as a bearing serving as a sound source. A CCD camera 18 or the like is connected as an image pickup device for presenting an image of the set position. Further, a camera drive unit 20 for directing the CCD camera 18 to the sound source position or zooming the position based on the actually specified sound source position of the abnormal sound is connected.
[0021]
Note that the microphones 12 used in the sound source search device 10 of the present embodiment are omnidirectional microphones, and a plurality of microphones are arranged in the sound source search region A at arbitrary intervals, preferably at equal intervals. In addition, it is desirable that a plurality of CCD cameras 18 are arranged at various places so as to be able to photograph all of the equipment such as facilities and apparatuses to be monitored in the sound source search area A.
[0022]
The processing unit 14 of the sound source search device 10 uses the microphones 12 to generate a sound waveform in the sound source search area A (the current sound waveform collected during the sound source search, and that no abnormal sound is generated as preparation). A sound (reference sound waveform) when a device in the sound source search area A is operating normally without generating an abnormal sound by the sound waveform collection unit 22 that collects a reference sound waveform to be collected at any time and the microphones 12. The reference sound waveform storage unit 24, which previously acquires and stores the current sound waveform, and compares the current sound waveform collected for each microphone 12 with the reference sound waveform of the corresponding microphone, which is stored in advance, and It includes a comparison unit 26 that obtains the difference and extracts only the abnormal sound waveform.
[0023]
Further, a residual information extracting unit (waveform processing unit) 28 for extracting residual information by removing only amplitude information from the extracted abnormal sound waveform, and a mutual relationship between residual information based on the current sound waveform collected by any two microphones. A time difference calculation unit 30 that calculates a correlation and calculates a time difference until an abnormal sound from the same sound source reaches each microphone for each combination of any two microphones 12 and a time difference calculated by a combination of each microphone 12 And a sound source specifying unit 32 for specifying a sound source position of the abnormal sound. The display control unit 34 for displaying the information on the specified sound source position on the display device 16 is connected to the information storage unit 36 storing the information on the sound source position. It is possible to provide specific related information of parts emitting noise, and information for eliminating abnormal sounds, for example, information for quickly and accurately performing maintenance work such as replacement or adjustment of parts. . This information should include, in detail, the model, drawing, specifications, etc. of the component, which are the details of the equipment component, history information such as the last replacement time of the component, life information, and the next scheduled replacement time. Is preferred.
[0024]
Further, a camera control unit 38 is connected to the sound source identification unit 32, and the camera drive unit 20 is operated to quickly turn the CCD camera 18 in the direction of the identified abnormal sound source. By displaying the portion where the abnormal sound is actually generated by the CCD camera 18, it is possible to increase the visual information for the manager and to perform maintenance work such as replacement or adjustment of parts more quickly. It is possible to do. Further, as described above, the maintenance procedure such as the replacement method of the target part is stored in the information storage unit 36, and the maintenance procedure and the like are displayed on the display device 16 together with the image acquired by the CCD camera 18. For example, even if an abnormal sound is generated from unfamiliar equipment, maintenance work can be performed quickly.
[0025]
A procedure for searching for a sound source of an abnormal sound by the sound source searching apparatus 10 configured as described above will be described below.
[0026]
First, as a preparation for the sound source search for abnormal sound, the sound waveform collection unit 22 uses each microphone 12 to include a sound when the device in the sound source search area A is operating normally, that is, completely includes the abnormal sound. The sound in the sound source search area A in the absence state is stored in the reference sound waveform storage unit 24 for each microphone 12 as a reference sound waveform. This collection of the reference sound waveform is preferably performed, for example, when all the devices are operated for the first time after the devices in the sound source search area A are completely inspected.
[0027]
When the sound source search device 10 actually performs the sound source search for the abnormal sound, the microphones 12 are caused to constantly collect the current sound waveform in the sound source search area A. The sound waveform collection unit 22 holds the current sound waveform at a predetermined sampling frequency, and transfers the data to the comparison unit 26 as data for each microphone 12 at a predetermined timing. At this time, the current sound waveform transferred to the comparison unit 26 is always emitted from a normal sound emitted from a device normally operating in the sound source search area A and from a device in which parts are worn or the like. It is a mixture of abnormal sounds. FIG. 2A schematically shows a current sound waveform in which a normal sound and an abnormal sound are mixed.
[0028]
On the other hand, when the current sound waveform collected by the microphone 12a is transferred, the comparison unit 26 reads out the reference sound waveform collected by the microphone 12a stored in the reference sound waveform storage unit 24, compares the two, and compares the two. Calculate the difference. FIG. 2B schematically shows the reference sound waveform stored in the reference sound waveform storage unit 24. FIG. 2C shows the difference between the current sound waveform and the reference sound waveform. As a result, the waveform of only the abnormal sound obtained is schematically shown. That is, the comparing unit 26 performs a process of extracting only the abnormal sound from the sounds collected by the microphone 12a. Since the extraction of a general abnormal sound is performed using a band filter or the like, the S / N ratio of only a predetermined frequency band is improved. However, as described above, the S / N ratio of all the frequency bands is compared with the reference waveform. / N ratio can be improved, and the accuracy of subsequent waveform processing can be improved.
[0029]
The comparison unit 26 also compares the current sound waveform collected by the microphone 12b from the sound waveform collection unit 22 with the reference sound waveform collected by the microphone 12b, and extracts the abnormal sound collected by the microphone 12b. Similarly, the same processing is performed on the current sound waveform collected by the microphone 12c to extract an abnormal sound.
[0030]
In the present embodiment, the abnormal sound means that the friction increases due to wear, shape deformation, lack of lubricant, etc. in the parts constituting the device, especially the motor part and the bearing part constituting the drive unit, and during normal operation. It mainly means sounds that do not occur, but also includes various sounds that occur due to malfunctions of the equipment.
[0031]
By the way, as shown in FIG. 3, when an abnormal sound is generated in a certain device (for example, the bearing 40), the distance to the microphone 12a at an arbitrary position is different from the distance to the microphone 12b. That is, a difference occurs in the time when the abnormal sound reaches the microphone 12. Conversely, if the time difference can be accurately grasped, the sound source position of the abnormal sound can be calculated by a mathematical method. When calculating this time difference, the time difference between the waveform collected by the microphone 12a and the waveform collected by the microphone 12b is generally calculated by cross-correlating the waveforms. However, since the sound is attenuated during propagation, At the microphones 12a and 12b located at different distances from the sound source, completely different waveforms are collected, and despite the waveforms being emitted from the same sound source, cross-correlation cannot be taken well. A high time difference cannot be calculated. This is because the attenuation of the amplitude of the waveform is included in the waveform as noise information.
[0032]
Therefore, in the present embodiment, the residual information extracting unit 28 performs a process of removing amplitude information from the abnormal sound waveform, and calculates a time difference from data having no noise component to improve accuracy and reliability. ing. Specifically, as shown in FIG. 4, the abnormal sound waveform a1 collected by the microphone 12a is Fourier-transformed, and separated into at least amplitude information b1 and phase information c1. Similarly, Fourier transform is performed on the abnormal sound waveform a2 collected by the microphone 12b to separate it into amplitude information b2 and phase information c2. As described above, the abnormal sound waveform a1 collected by the microphone 12a and the abnormal sound waveform a2 collected by the microphone 12b are abnormal sounds emitted from the same sound source. Are recognized as different waveforms.
[0033]
Further, the residual information extracting unit 28 combines only the separated phase information c1 and phase information c2, creates a combined waveform d, and provides the combined waveform d to the time difference calculating unit 30. The time difference calculation unit 30 performs an inverse Fourier transform on the obtained composite waveform d to create a correlation waveform e. In the correlation waveform e created here, as shown in FIG. 5A, the peak of the frequency corresponds to the time difference t1 when the waveforms emitted from the same sound source are collected by the two microphones 12a and 12b. Appears. That is, it is detected that the arrival time of the abnormal sound has a time difference of, for example, t1 = 10 ms between the two microphones 12a and 12b.
[0034]
As described above, the cross-correlation of abnormal sounds collected by the microphones 12a and 12b is improved by using only the phase information of the abnormal sound waveform and eliminating the influence of attenuation due to the difference in the distance between the sound source and the microphone 12. And the time difference can be calculated clearly.
[0035]
Similarly, the residual information extraction unit 28 and the time difference calculation unit 30 perform processing on the abnormal sound waveform acquired by the microphones 12b and 12c, and perform the processing for the time difference t2 (see FIG. 5B) and the time difference t2 between the microphone 12c and the microphone 12a. Processing is performed on the acquired abnormal sound waveform to calculate a time difference t3 (see FIG. 5C), and the calculated time differences t1, t2, and t3 are supplied to the sound source identification unit 32.
[0036]
When the sound source identification unit 32 acquires the time difference t1, as shown in FIG. 6, the arrival time m of the sound between the microphone 12a and the imaginary sound source 42 and the arrival time n of the sound between the microphone 12b and the sound source 42 , A rotational hyperbolic surface S1 connecting candidate points p1, p2, p3,... That is, the sound source 42 that can form the time difference t1 with respect to the microphones 12a and 12b is located on one of the round-trip transfer curved surfaces S1.
[0037]
Similarly, the sound source identification unit 32 creates a rotational hyperbolic surface S2 based on the time difference t2 regarding the microphones 12b and 12c and a rotational hyperbolic surface S3 based on the time difference t3 regarding the microphones 12c and 12a. Of course, the sound source 42 is located on one of the rotational hyperbolic surfaces S2 and S3. Therefore, as shown in FIG. 7, the sound source identification unit 32 determines the position of the sound source 44 that emits the actual abnormal sound by searching for the position where the created rotational hyperboloid surfaces S1, S2, and S3 intersect. Can be.
[0038]
When the position of the sound source 44 emitting the actual abnormal sound is determined on the XYZ coordinate axes, the sound source identification unit 32 transfers the determined coordinates of the sound source 44 emitting the abnormal sound to the display control unit 34. The sound source identification unit 32 searches the information storage unit 36 based on the coordinates of the sound source 44, and calls up information on a device existing at the corresponding coordinates, and further information on a part of the device that matches the coordinates of the sound source 44. For example, as shown in FIG. 8, “the main shaft bearing B of the facility A: model = OO” is displayed on the display device 16. Also, at this time, the CAD drawing of the part constituting the sound source of the abnormal sound searched is retrieved and displayed from the information storage unit 36 or another database, the specification of the part, the life curve, the replacement history, the next scheduled replacement time, etc. May also be displayed. Further, a procedure for replacing parts constituting the sound source of the abnormal sound may be indicated.
[0039]
In the present embodiment, if the coordinates of the sound source 44 of the abnormal sound can be specified by the sound source specifying unit 32, the coordinates are supplied to the camera control unit 38. The camera control unit 38 drives the camera drive unit 20 of the CCD camera 18 near the coordinates based on the acquired coordinates of the sound source 44 of the abnormal sound, and moves the CCD camera 18 to the coordinates, that is, the abnormal sound. The video is displayed on the display device 16 toward the emitting device. Note that the image to be displayed may be only a part that is actually emitting an abnormal sound, or may be an entire target device, a specific portion of the device, or a parallel display of each image. At this time, as shown in FIG. 8, the information may be displayed together with the various types of information described above.
[0040]
In this way, by displaying various kinds of information on the sound source position of the abnormal sound, it is possible to clearly present the sound source of the abnormal sound to the administrator, and to efficiently perform maintenance work after recognizing the cause of the abnormal sound. Can be done Further, a maintenance work can be performed more smoothly by mutually referencing CAD drawings of parts, replacement methods, images of actual parts, and the like. Of course, the display content shown in FIG. 8 is an example, and it is desirable to appropriately increase or decrease the types of information that can be searched based on the coordinates of the sound source 44 of the abnormal sound.
[0041]
By the way, in the sound source search area A, since a plurality of devices are usually operating at the same time, an abnormal sound may be emitted from a plurality of devices at the same time. That is, a plurality of abnormal sound sources may exist at the same time. Also in this case, basically, the sound source position of each abnormal sound can be specified by performing the same processing as in the case of one abnormal sound described above.
[0042]
For example, when three abnormal sound sources exist, the current sound waveform collected by the microphone 12a has a waveform including three abnormal sounds P, Q, and R in the normal sound. Similarly, the waveforms collected by the microphones 12b and 12c are also waveforms including three abnormal sounds P, Q, and R in the normal sound.
[0043]
As a result, the waveform calculated by the comparison process in the comparison unit 26 also includes three abnormal sound waveforms. However, at this time, the order in which the three abnormal sounds P, Q, and R appear differs depending on the distance to each microphone 12.
[0044]
The residual information extraction unit 28 obtains a plurality of pieces of phase information corresponding to the abnormal sounds P, Q, and R by performing a Fourier transform on a waveform including the abnormal sounds P, Q, and R. Then, a synthesized waveform is created based on the phase information acquired by the microphones 12a and 12b, and the time difference calculation unit 30 performs an inverse Fourier transform to obtain a waveform emitted from a different sound source as shown in FIG. When collected by the two microphones 12a and 12b, a frequency peak appears at a position corresponding to the time difference between the same sound sources. In this case, peak values appear at three positions. Similarly, as shown in FIGS. 9B and 9C, frequency peaks appear at positions corresponding to the time differences when the microphones 12b and 12c and the microphones 12c and 12a collect. Since the distances from the sound source to the microphones 12 are individually different, the times at which the abnormal sounds P, Q, and R reach the microphones 12 are also different. As a result, as shown in FIGS. 9A to 9C, the appearance order and frequency of the peak value p relating to the abnormal sound P, the peak value q relating to the abnormal sound Q, and the peak value r relating to the abnormal sound R are individually determined. Different. However, by referring to the phase information of each abnormal sound P, Q, R, which peak value p, q, r appearing in FIG. Can be recognized.
[0045]
The sound source identification unit 32 extracts a time difference based on a peak value p related to an abnormal sound of the same sound source, for example, the abnormal sound P, from the time difference obtained by the combination of the microphones 12, and is similar to that described with reference to FIGS. 6 and 7. Next, the rotational hyperbolic surfaces S1, S2, and S3 for the abnormal sound P are created, and the sound source position of the abnormal sound P is specified. Similarly, a time difference based on the peak value of the abnormal sound Q is extracted, hyperboloids of revolution S1, S2, and S3 are created, and the sound source position of the abnormal sound Q is specified. The same applies to the abnormal sound R.
[0046]
As described above, in the present embodiment, even when a plurality of sound sources generating abnormal sounds are present in the sound source search area A, the position of the sound source can be specified with high accuracy.
[0047]
Also in this case, after the sound source of the abnormal sound can be identified, the information related to the sound source is individually displayed as shown in FIG. 8 so that even if there are a plurality of sound sources, the manager is not confused. Appropriate measures can be taken promptly.
[0048]
Note that the configuration illustrated in FIG. 1 is an example, and any configuration is optional as long as it has a function of performing the same processing, and an effect similar to that of the present embodiment can be obtained. For example, it may be realized by using the processing function of a personal computer, or may be realized by improving an existing waveform analyzer.
[0049]
【The invention's effect】
According to the present invention, since cross-correlation is performed by using residual information excluding amplitude information that causes noise, recognition accuracy of mutual information correlation is improved, and abnormal sounds from the same sound source reach each microphone. It is possible to obtain the time difference until the calculation with high accuracy. As a result, the position of the sound source of the abnormal sound can be accurately specified. In this case, even when the number of abnormal sound sources is one or more than one, the correlation between individual abnormal sounds can be accurately grasped, so that an accurate sound source search can be performed.
[Brief description of the drawings]
FIG. 1 is a configuration block diagram illustrating a configuration concept of a sound source search device that realizes a method of searching for a sound source of an abnormal sound according to an embodiment of the present invention.
FIG. 2 is an explanatory diagram illustrating extraction of an abnormal sound waveform from a collected current sound waveform in the abnormal sound source searching method according to the embodiment of the present invention.
FIG. 3 is an explanatory diagram for explaining that a time difference occurs between sounds reaching arbitrary two microphones in the abnormal sound source searching method according to the embodiment of the present invention.
FIG. 4 is an explanatory diagram illustrating a procedure for calculating a time difference between sounds reaching two arbitrary microphones in the method for searching for a sound source of an abnormal sound according to the embodiment of the present invention.
FIG. 5 is an explanatory diagram for explaining a time difference obtained by a combination of two arbitrary microphones in the abnormal sound source searching method according to the embodiment of the present invention.
FIG. 6 illustrates a procedure for obtaining a rotational hyperbolic surface where a sound source may exist based on a time difference between sounds reaching two arbitrary microphones in the method for searching for a sound source of an abnormal sound according to the embodiment of the present invention. FIG.
FIG. 7 is an explanatory diagram illustrating that a sound source is specified by three rotational hyperbolic planes obtained by a combination of two arbitrary microphones in the method for searching for an abnormal sound source according to the embodiment of the present invention.
FIG. 8 is an explanatory diagram illustrating a display example of information on a sound source specified by the abnormal sound source search method according to the embodiment of the present invention.
FIG. 9 is a diagram illustrating a method for detecting a sound source of an abnormal sound according to an embodiment of the present invention, wherein when a plurality of (for example, three) abnormal sound sources exist, a time difference obtained by a combination of two arbitrary microphones; FIG.
[Explanation of symbols]
Reference Signs List 10 sound source search device, 12, 12a, 12b, 12c microphone, 14 processing unit, 16 display device, 18 CCD camera, 20 camera driving unit, 22 sound waveform collection unit, 24 reference sound waveform storage unit, 26 comparison unit, 28 residue Information extraction unit, 30 time difference calculation unit, 32 sound source identification unit, 34 display control unit, 36 information storage unit, 38 camera control unit.

Claims (6)

Collecting the current sound waveform that is generated when the device is operating in the sound source search area using three or more microphones arranged at different positions in the sound source search area;
Extracting a difference between the reference sound waveform collected by each microphone and the current sound waveform when the device is operating normally in the sound source search area, and extracting an abnormal sound waveform;
A step of eliminating only the amplitude information of the extracted abnormal sound waveform and obtaining residual information not including the amplitude information,
Cross-correlating residual information based on the current sound waveform collected by any two microphones to determine a time difference until an abnormal sound from the same sound source reaches each microphone;
Identifying a sound source position of an abnormal sound based on a time difference obtained by a combination of the microphones;
A method for searching for a sound source of an abnormal sound, comprising: The method of claim 1, wherein
Based on the identified sound source position, identifying a corresponding position of the device present in the sound source search area,
Reading and presenting information on the device at the identified corresponding position from the storage unit;
A method for searching for a sound source of an abnormal sound, comprising: The method according to claim 1 or claim 2,
Based on the specified sound source position, a step of controlling imaging means for acquiring an image of a corresponding position of the sound source in the sound source search area,
Presenting the acquired image;
A method for searching for a sound source of an abnormal sound, comprising: At least three microphones arranged at different positions in the sound source search area and collecting current sound waveforms generated when the device is operating in the sound source search area;
A reference sound storage unit that stores a reference sound waveform generated when the device is operating normally in the sound source search area,
For each microphone, a comparison unit that compares the current sound waveform and the reference sound waveform collected at the same position and extracts an abnormal sound waveform,
A waveform processing unit that obtains residual information by excluding only amplitude information from each extracted abnormal sound waveform,
A time difference calculator for collecting cross-correlation of residual information based on the current sound waveform collected by any two microphones and obtaining a time difference between abnormal sounds from the same sound source reaching each microphone;
A sound source identification unit that identifies a sound source position of an abnormal sound based on a time difference obtained by a combination of each microphone;
A sound source locating device for abnormal sound, comprising: The device according to claim 4,
A search unit that searches for information on the corresponding position of the device in the sound source search area corresponding to the specified sound source position;
A presentation unit for presenting the searched information;
A sound source locating device for abnormal sound, comprising: The device according to claim 4 or 5,
An imaging unit that acquires an image of a sound source corresponding position in the sound source search area based on the specified sound source position,
A presentation unit for presenting a captured image;
A sound source locating device for abnormal sound, comprising:

Images