Environmental sound monitoring device, environmental sound monitoring method, and program

Abstract

To simply and accurately estimate the position of a sound source that generates environmental sound at a construction site.SOLUTION: An environmental sound monitoring device 100 includes a sound source position estimation unit 113 which calculates a sound source estimation direction vector for each of microphone arrays 200 on the basis of audio signals from the plurality of microphone arrays 200, calculates planar position coordinates of all intersection points of the sound source estimation direction vectors obtained for the respective microphone arrays 200, classifies the plurality of intersection points into a predetermined number of clusters, calculates a centroid of the intersection points belonging to each cluster, and determines a sound source estimation position on the basis of an index indicating the degree of concentration of the intersection points with respect to the centroid for each cluster.SELECTED DRAWING: Figure 4

Description

[Technical Field]

[0001] The present invention relates to an environmental sound monitoring device, an environmental sound monitoring method, and a program. [Background technology]

[0002] At construction sites, sounds generated by heavy machinery and various types of work (hereinafter referred to as "environmental sounds") can have an impact on the living environment around the site. Environmental standards and the Noise Control Act set standard values ​​at site boundaries, and specific construction work is required to adhere to these standard values.

[0003] Currently, to comply with the standard values, omnidirectional microphones are installed at the boundaries of construction sites to monitor the environmental sound levels at the boundaries. However, omnidirectional microphones only measure sound pressure levels and cannot determine the location of the sound source. Therefore, a method is used in which multiple microphone arrays, consisting of multiple microphones placed in close proximity to each other, are used to estimate the direction of sound arrival and identify the sound source location by superimposing the estimated vectors of the direction of arrival. Figure 1 shows a schematic diagram of a situation in which sound source location is estimated using multiple microphone arrays. For simplicity, in Figure 1, the positions of two sound sources SS1 and SS2 within a rectangular site F are estimated using three microphone arrays 200. The process of estimating the direction of sound arrival using each microphone array 200 is described, for example, in Futoshi Asano, "Forest of Knowledge, Group 2 (Image, Sound, Language), Part 6 (Acoustic Signal Processing), Chapter 3: Sound Source Localization," [online], November 2011, Institute of Electronics, Information and Communication Engineers, [searched February 21, 2024], Internet.<https: / / www.ieice-hbkb.org / files / 02 / 02gun_06hen_03.pdf> As shown in Fig. 1, the positions of two sound sources SS1 and SS2 can be identified as the intersection of vectors indicating the sound arrival directions obtained by three microphone arrays 200 (hereinafter referred to as "sound source position estimation vectors").

[0004] However, this method also detects an intersection P of vectors resulting from sounds received from different sound sources at each microphone array 200. Although this intersection P is not related to sound source estimation, it cannot be distinguished from an intersection indicating the sound source position as it is.

[0005] Regarding this, for example, in Non-Patent Document 1, the spectrum of the incoming sound source is calculated for each combination of two direction vectors for all intersections of the obtained sound source direction estimation vectors, clustering is performed from the viewpoint of the similarity of the spectra, and it is determined from the similarity whether the intersection is an actual sound source or not. [Prior art documents] [Non-patent literature]

[0006] [Non-Patent Document 1] Daniel Gabriel, et.al., "Iterative Outlier Removal Method Using In-Cluster Variance Changes in Multi-Microphone Array Sound Source Localization", The Proceedings of The 79th National Convention of IPSJ, pp. 2-229-230, Nagoya University, Nagoya, Mar. 16-18,2017 Summary of the Invention [Problem to be solved by the invention]

[0007] However, the method proposed in Non-Patent Document 1 has a problem in that the calculation load required for spectrum calculation increases as the number of combinations of direction vectors increases, resulting in a longer calculation time. This becomes a problem particularly when attempting to repeatedly perform calculations for sound source estimation in order to continuously monitor changes in environmental sounds over time.

[0008] An object of the present invention is to provide an environmental sound monitoring device, an environmental sound monitoring method, and a program that can easily and accurately estimate the position of a sound source that generates environmental sound at a construction site. [Means for solving the problem]

[0009] One aspect of the present invention is an environmental sound monitoring device that includes a sound source position estimation unit that calculates a sound source estimation direction vector for each of a plurality of sound collection units based on audio signals from the sound collection units, each of which is equipped with a plurality of sound collection means, calculates planar position coordinates of all intersections of the sound source estimation direction vectors obtained for each of the sound collection units, classifies the intersections into a predetermined number of clusters, calculates the center of gravity of the intersections to which each cluster belongs, and determines an estimated sound source position for each of the clusters based on an index that indicates the density of the intersections relative to the center of gravity.

[0010] The environmental sound monitoring device may further include a sound source analysis unit that calculates the acoustic power level of each sound source based on the determined estimated sound source position, the position of each sound collection unit, and the sound pressure level measured by each sound collection unit.

[0011] For each of the clusters, the sum of the reciprocals of the distances between the center of gravity and each of the intersections may be calculated, and the center of gravity of the cluster for which the sum is greater than a predetermined threshold may be identified as the estimated position of the sound source.

[0012] For each cluster, the sum of the reciprocals of the distances between the center of gravity and each of the intersections may be calculated, the sum values ​​may be clustered into two clusters, and the center of gravity of the cluster with the larger average value of the two obtained clusters may be identified as the sound source estimation position.

[0013] Another aspect of the present invention is an environmental sound monitoring method in which a computer calculates a sound source estimation direction vector for each of a plurality of sound collection units based on audio signals from the sound collection units, each of which has a plurality of sound collection means, calculates planar position coordinates of all intersections of the sound source estimation direction vectors determined for each of the sound collection units, classifies the intersections into a predetermined number of clusters, calculates the center of gravity of the intersections to which each cluster belongs, and determines an estimated sound source position for each cluster based on an index indicating the density of the intersections relative to the center of gravity. A program that causes a computer to execute this environmental sound monitoring method is also one aspect of the present invention. [Effects of the Invention]

[0014] According to the present invention, it is possible to provide an environmental sound monitoring device, an environmental sound monitoring method, and a program that can easily and accurately estimate the position of a sound source that generates environmental sound at a construction site. [Brief explanation of the drawings]

[0015] [Figure 1] FIG. 1 is a schematic diagram illustrating the principle of sound source localization. [Figure 2] 1 is a schematic diagram illustrating an example of the overall configuration of an environmental sound monitoring system 1 according to an embodiment of the present invention. [Figure 3] 1 is a block diagram illustrating the configuration of an environmental sound monitoring device 100 according to an embodiment of the present invention. [Figure 4] 1 is a block diagram illustrating an example of the configuration of a processing unit 110 included in the environmental sound monitoring device 100. FIG. [Figure 5] 1 is a flowchart showing an example of data processing executed by the environmental sound monitoring device 100. [Figure 6] FIG. 10 is an explanatory diagram of a clustering process for the purpose of estimating a sound source position. [Figure 7] 10 is a histogram illustrating the results of a clustering process. [Figure 8] 1 is a schematic diagram showing an example of output of sound source positions and sound power level calculation results by the environmental sound monitoring device 100. FIG. DETAILED DESCRIPTION OF THE INVENTION

[0016] Hereinafter, the present invention will be described based on an embodiment thereof with reference to the accompanying drawings.

[0017] <Environmental Sound Monitoring Method in This Embodiment> First, an environmental sound monitoring method according to one embodiment of the present invention will be described. As described with reference to FIG. 1 , in this embodiment, multiple microphone arrays, each including multiple microphones, are installed around the boundary of the construction site F to record various environmental sounds generated at the construction site F. For simplicity's sake, FIG. 1 illustrates a configuration in which three microphone arrays 200 collect environmental sounds propagating from two sound sources at the construction site F. When multiple microphone arrays are installed to monitor environmental sounds at the construction site F, the sound source direction estimation vectors for each microphone array's sound source increase as the number of sound sources increases, and the number of intersections between the sound source direction estimation vectors also increases accordingly. As described above, if the location of a sound source is not known in advance, it is impossible to determine which of these intersections indicates the estimated location of the sound source. Furthermore, the intersections estimated to be the sound source will vary in position relative to the sound source due to noise other than the sound from the sound source.

[0018] Therefore, in this invention, based on the knowledge that the sound source is estimated to be located near a point where the density of intersections of multiple sound source direction estimation vectors detected based on sounds collected by multiple microphone arrays within the construction site F is high, a method has been established in which the multiple intersections are clustered and points where the intersections are relatively densely located are identified to estimate the location of the sound source.

[0019] The configuration for realizing the above-described sound source position estimation method, as well as its actions and effects, will be described below with reference to the embodiments.

[0020] <Configuration example of environmental sound monitoring system> First, an example of the configuration of an environmental sound monitoring system 1 according to this embodiment will be described with reference to Fig. 2. Fig. 2 is a block diagram showing an example of the configuration of the environmental sound monitoring system 1. The environmental sound monitoring system 1 includes an environmental sound monitoring device 100 and a plurality of microphone arrays 200. The environmental sound monitoring device 100 is a computer equipped with a processor and memory, and has the functions of receiving audio signals from each microphone array as input, calculating the direction of arrival of sound propagating from a sound source, estimating the position of the sound source based on the calculated direction of arrival, calculating the acoustic power level of the sound source at the estimated position, and outputting the calculation results. An example of the configuration of the environmental sound monitoring device 100 will be described later.

[0021] The audio signals output from the microphones 210 provided in each microphone array 200 can be transmitted as analog signals via a wire to the environmental sound monitoring device 100. Alternatively, the audio signals can be converted into digital signals in each microphone array 200 and transmitted via a wire or wirelessly to the environmental sound monitoring device 100. Wireless transmission can be performed, for example, by Wi-Fi.

[0022] The number of microphones 210 provided in each microphone array 200 can be determined taking into consideration factors such as the accuracy of estimating the sound arrival direction and the calculation load for estimating the sound arrival direction.

[0023] The number of microphone arrays 200 to be installed at a construction site where environmental sound monitoring is performed and their placement at the site can be determined based on factors such as the size of the construction site and the location of the anticipated sound source. As an example, it is preferable to place the microphone array 200 so as to surround the anticipated sound source. Furthermore, if the shape of the construction site is polygonal, it is preferable to place the microphone array 200 as close to the sides of the polygon as possible, and to avoid the vertices of the polygon and their vicinity. This is because if the microphone array 200 is placed near the vertices, the direction of the sound collected by the microphone array will be narrower than if it were placed on the sides, and the direction of the obtained sound source localization vector will be limited. It is preferable that sounds arrive from a wide range of sound sources in the microphone array, and the direction of the sound source localization vector will be as wide as possible. This is because the intersection angle between the sound source localization vectors becomes closer to a right angle, which is expected to improve the accuracy of sound source localization.

[0024] <Environmental Sound Monitoring Device 100> Next, the configuration of the environmental sound monitoring device 100 will be described. Fig. 3 is a block diagram illustrating an example of the hardware configuration of the environmental sound monitoring device 100 in one embodiment of the present invention. As shown in Fig. 3, the environmental sound monitoring device 100 in this embodiment includes a processing unit 110, a storage unit 120, an input unit 130, an output unit 140, and a communication unit 150.

[0025] The environmental sound monitoring device 100 can be realized as a general computer of various types, including a desktop type, a portable type, etc. Alternatively, a program for realizing the functions of the environmental sound monitoring device 100 may be installed in a server computer connected to a network, and the environmental sound monitoring device 100 may be configured to be available from a client computer via the network.

[0026] The hardware elements of the environmental sound monitoring device 100 illustrated in Fig. 3 will now be described. The processing unit 110 is configured by a processor that performs various types of calculations, control, and other processing required for the operation of the environmental sound monitoring device 100. The processor that configures the processing unit 110 is, for example, a CPU (Central Processing Unit), MPU (Micro Processing Unit), SoC (System on a Chip), DSP (Digital Signal Processor), GPU (Graphics Processing Unit), ASIC (Application Specific Integrated Circuit), PLD (Programmable Logic Device), FPGA (Field-Programmable Gate Array), or a combination thereof. The processing unit 110 may also be a combination of such a processor with a hardware accelerator, etc.

[0027] The storage unit 120 provides a storage area for storing programs to be executed by the processing unit 110, a storage area as a work area temporarily used in performing various arithmetic processing, and a storage area for storing parameter data etc. required for executing the programs. The storage unit 120 is configured by, for example, a ROM (Read Only Memory) which is a nonvolatile memory, a RAM (Random Access Memory) which is a volatile memory, a semiconductor memory, a hard disk drive etc.

[0028] Input unit 130 includes an interface circuit that receives input data input by user operation from various input devices such as a keyboard, mouse, or touch panel, and an audio signal from microphone array 200. Output unit 140 includes an interface circuit that outputs data signals to output devices such as a display that displays images and a printer for output printing, and outputs still image / movie image data, text data, and the like.

[0029] The communication unit 150 includes an interface circuit that controls communication between the environmental sound monitoring device 100 and external devices such as a management server via various communication networks such as a LAN (Local Area Network), and an interface circuit that controls communication with the microphone array 200 when audio signals from the microphone array 200 are transmitted as digital signals. Note that power is supplied to the microphone array 200 and the environmental sound monitoring device 100 from a temporary power source provided at the construction site or a commercial AC power source.

[0030] <Functions of the Environmental Sound Monitoring Device 100> Next, functions realized by the processing unit 110 included in the environmental sound monitoring device 100 will be described. FIG. 4 is a block diagram illustrating functions realized by the processing unit 110 included in the environmental sound monitoring device 100 according to one embodiment of the present invention. As shown in FIG. 4, the processing unit 110 includes an input data receiving unit 111, an arrival direction calculation unit 112, a sound source position estimation unit 113, an acoustic power level calculation unit 114, a calculation result output unit 115, an input control unit 116, and an output control unit 117. These functional units are realized, for example, as a program executed by the processing unit 110. The program may be realized in any form, but as an example, they may be realized as a program that runs on an appropriate operating system (OS) implemented in the environmental sound monitoring device 100.

[0031] The input data receiving unit 111 has a function of receiving, via the input unit 130, input data used by a program running on the environmental sound monitoring device 100. The data received by the input data receiving unit 111 is passed to the arrival direction calculation unit 112, the sound source position estimation unit 113, and the sound power level calculation unit 114 for use.

[0032] The arrival direction calculation unit 112 calculates the arrival direction of the sound captured by the microphones 210 of the microphone array 200 using a known sound source localization method such as an acoustic intensity method or a beamforming method, based on the difference in sound reception timing between the microphones 210 constituting the microphone array 200. Specifically, the arrival direction calculation unit 112 is realized as a program or the like that processes data using an algorithm of the MUSIC (MUltiple SIgnal Classification) method, for example. A sound source position estimation vector is calculated for each microphone array 200 based on the arrival angle calculated here and the installation position of the microphone array 200 on a plane coordinate set for the construction site where the microphone array 200 is installed. Data of the calculated sound source position estimation vector is passed to the sound source position estimation unit 113. The arrival direction calculation process may be executed by a calculation device provided in the microphone array 200.

[0033] The sound source position estimation unit 113 performs a sound source position estimation process based on the sound source position estimation vector obtained for each microphone array 200 input from the arrival direction calculation unit 112. The sound source position estimation vector basically indicates the direction of the sound source position from the sound sources of various environmental sounds at a construction site to each microphone array 200. However, at a construction site, many acoustic components that become noise when estimating the sound source position, such as noise from various work noises generated in addition to specific sound sources, are included. Therefore, the sound source position estimation unit 113 of this embodiment performs a calculation process to remove these noise components and accurately determine the position of the sound source present at the construction site. The calculation process will be described below.

[0034] (1) Sound source localization processing policy As described with reference to FIG. 1 , the position of a sound source to be determined should theoretically be found as an intersection of sound source position estimation vectors obtained by the microphone array 200. However, in reality, due to sounds other than the sound source to be determined, such as inherent noise components contained in the microphone, the sound source position estimation vectors may include vectors that do not correctly indicate the direction of the sound source position. Therefore, not all intersections of the obtained sound source position estimation vectors indicate the sound source position, and due to calculation errors of the sound source position estimation vectors, the intersections of the sound source position estimation vectors do not necessarily correlate with the sound source position. On the other hand, it is considered that the majority of sound source position estimation vectors indicate the sound source position or at least its vicinity. Therefore, the sound source estimation position is expected to exist in an area where the intersections of the sound source position estimation vectors are densely concentrated. Based on the above reasoning, the present invention has arrived at a policy that enables a sound source to be meaningfully estimated to be located at a location where the intersections of all sound source position estimation vectors are densely concentrated at a certain level or more in the distribution of the intersections of all sound source position estimation vectors.

[0035] (2) Sound source location estimation process The data processing for realizing the above policy will be described below. (a) The intersection of all sound source position estimation vectors obtained by the microphone array 200 is found. An appropriate origin is placed in advance for the construction site where environmental sound monitoring will be performed, and plane coordinates (x-y coordinates) are set. Based on the position coordinates of each microphone array 200 and the obtained sound source position estimation vector, the position coordinates of all intersections are calculated.

[0036] (b) All intersections are classified into an appropriate number of clusters, and the center of gravity of each cluster is determined from the position coordinates of the intersections included in each cluster. The number of clusters is set to be greater than the number of expected sound sources. As a clustering method, for example, the k-means method can be used, but other clustering methods may also be used. The center of gravity of the intersections determined for each cluster is assumed to be the position of the sound source.

[0037] (c) For the center of gravity of the intersections found for each cluster, the influence of the intersections that are not considered to contribute to estimating the sound source position is reduced as much as possible, and an operation is performed to bring the center of gravity of the intersections asymptotically closer to the sound source estimated position. Specifically, in the process of repeatedly performing the calculation to find the center of gravity of the intersections, the influence of the intersections farther from the found center of gravity position is reduced in the calculation to calculate the next center of gravity position. When the difference between the found latest center of gravity position and the previous center of gravity position falls below a predetermined value, it is determined that the center of gravity position is sufficiently close to the hypothetical sound source position. In this way, the center of gravity position is found as the hypothetical sound source position for each cluster.

[0038] (d) Regarding the hypothetical estimated sound source positions obtained for each cluster, which one indicates the actual sound source position is identified using the density of intersections related to the hypothetical estimated sound source positions as an index. As an index, for example, the sum of the reciprocals of the distances between the hypothetical estimated sound source position and each intersection can be used. The larger this index, the more densely the intersections related to the hypothetical estimated sound source positions in the cluster are said to be. Therefore, this index is classified into clusters using an appropriate method, and the center of gravity of the cluster identified as having a relatively large index is determined to be the actual estimated sound source position.

[0039] In this way, the position where the actual sound source exists can be estimated with high accuracy based on a large number of intersections of sound source position estimation vectors obtained by a plurality of microphone arrays.

[0040] Returning to the functions of the processing unit 110, the sound power level calculation unit 114 has the function of calculating the sound power level of each sound source based on the position on the planar coordinates of the construction site estimated by the sound source position estimation unit 113, the position of each microphone array, and the sound pressure level corresponding to the sound source position estimation vector indicating the actual sound source estimated position measured by each microphone array.

[0041] The calculation result output unit 115 has a function of outputting the estimated sound source position calculated by the sound source position estimation unit 113 and the sound power level of each sound source calculated by the sound power level calculation unit 114 to the output control unit 117.

[0042] The input control unit 116 executes a process for accepting data input to an input device that constitutes the input unit 130 .

[0043] The output control unit 117 executes a process of outputting data to an output device constituting the output unit 140. For example, the output control unit 117 executes a process of generating display data or the like based on the output data generated by the calculation result output unit 115 and displaying the data on a screen provided in the output unit 140.

[0044] In this way, the environmental sound monitoring device 100 can obtain the sound source position on a predetermined plane coordinate determined by a multiple microphone array and the acoustic power level of each sound source, and can continuously monitor the status of environmental sounds that may become noise and are generated from construction sites.

[0045] <Data Processing by Environmental Sound Monitoring Device 100> Next, a process for estimating a sound source position by the environmental sound monitoring device 100 of this embodiment and a process for calculating the sound power level of each estimated sound source will be described. Fig. 5 shows a flowchart illustrating an example of a data processing flow executed by the environmental sound monitoring device 100 provided in the environmental sound monitoring system 1 of this embodiment. The sound source position estimation process and sound power level calculation process in this embodiment are started by supplying power to a computer functioning as the environmental sound monitoring device 100 and to each connected microphone array 200, and by starting a program that realizes the functions of the environmental sound monitoring device 100. Power is supplied to each microphone array 200 from the connected environmental sound monitoring device 100, or from a temporary on-site power source, etc.

[0046] When the program of the environmental sound monitoring device 100 is started, first, in step S11, environmental sounds are collected by the microphones 210 constituting the microphone array 200 and input to the arrival direction calculation unit 112. The audio signals transmitted from each microphone array 200 are identified by the connection channel of the microphone 210 that collected the sound. If the audio signals are converted into digital signals in advance in the microphone array 200, an ID identifying each microphone 210 may be included. This allows the arrival direction calculation unit 112 to calculate the arrival time difference of the environmental sounds input from each microphone 210, the sound source positions of which are to be estimated. Note that the audio signals collected by the microphone array 200 may be input directly to the arrival direction calculation unit 112 for real-time processing, or the collected audio signals may be temporarily stored in the storage unit 120 or the like, and later, as needed, the audio signals may be read out and input to the arrival direction calculation unit 112 for offline processing.

[0047] In step S12, arrival direction calculation unit 112 calculates the arrival time difference between microphones 210 of microphone array 200 for the input environmental sound, and based on that, calculates the arrival direction of the sound relative to microphone array 200. Arrival direction calculation unit 112 further calculates a sound source estimation direction vector for each microphone array 200 using the arrival direction of the sound and the installation position coordinates of each microphone array 200. Note that the arrival direction of the sound may also be calculated using other processing methods such as an acoustic intensity method.

[0048] In step S13, the sound source position estimation unit 113 finds all intersections (xi, yi) on the xy coordinate plane of the multiple sound source estimation direction vectors for each microphone array 200 found by the arrival direction calculation unit 112, and performs non-hierarchical clustering on the two-dimensional coordinate plane. This non-hierarchical clustering classifies the intersections into multiple clusters using an appropriately selected clustering method. As mentioned above, the k-means method is a common clustering method. For example, in the k-means method, it is sufficient to set a larger number of clusters in advance, assuming that the number of sound sources will be classified into clusters that is greater than the number of sound sources expected in advance.

[0049] Other methods that may be used include K-medoids, spectral clustering, Density-Based Spatial Clustering of Applications with Noise (DBSCAN), which specifies the minimum number of intersections constituting a cluster and the size of the cluster region (distance in this case), and X-means, which converges the number of clusters by repeating the k-means method. Fig. 6 schematically shows a state in which all intersections (indicated by circles) of sound source position estimation vectors on a coordinate plane are classified into five clusters C1 to C5. Note that the schematic diagram shown in Fig. 6 is a fictitious diagram to facilitate understanding of this embodiment and does not illustrate actual clustering results.

[0050] In step S14, the sound source position estimation unit 113 executes the following processing steps from step S15 to step S18 for all clusters.

[0051] First, in step S15, the sound source position estimation unit 113 calculates the coordinates of the center of gravity (xp, yp) of the intersection group (xi, yi) included in each cluster. Here, i means the i-th intersection, where i = 1 to Nj. Nj is the number of intersections included in the j-th cluster. In Fig. 6, the centers of gravity of the intersections of each of the clusters C1 to C5 are indicated by + signs. The initial value of the center of gravity G of each cluster is the average value of the coordinates of all the intersections included in the cluster.

[0052] In step S16, the sound source position estimation unit 113 weights each intersection by the inverse of the distance between each intersection and the center of gravity, and calculates the center of gravity again. Specifically, the distance li between each intersection and the center of gravity G is calculated as follows:

number

number

number

[0053] FIG. 6 shows a schematic diagram of cluster C5 in which the influence of the intersections far from the center of gravity is reduced and the center of gravity position (+ mark) approaches the distance δ from the actual sound source position (★ mark).

[0054] In step S17, the sound source position estimation unit 113 determines whether the distance between the previously calculated pre-update center of gravity position and the updated center of gravity position is equal to or less than a predetermined value. If it is determined that the distance between the pre-update center of gravity position and the updated center of gravity position exceeds the predetermined value (step S17, No), the sound source position estimation unit 113 returns to step S16 and executes the process of updating the center of gravity position. If it is determined that the distance between the pre-update center of gravity position and the updated center of gravity position is equal to or less than the predetermined value (step S17, Yes), the sound source position estimation unit 113 determines in step S18 that the sound source estimated position has been obtained as the center of gravity position for the corresponding cluster. As described above, the process of obtaining this tentative sound source estimated position as the center of gravity position is executed for each cluster (step S19). Note that the predetermined value for the above-mentioned determination can be set through an experiment of sound source position estimation using an actual sound source.

[0055] In step S20, the sound source position estimation unit 113 finds the center of gravity of each cluster and the sum of the reciprocals of the intersection distances for each cluster. The sum dj of the reciprocals of the intersection distances can be expressed by the following equation, where j represents the cluster number.

number

[0056] As described above, this embodiment employs a policy that assumes that the number of intersections of sound source direction estimation vectors is large near a sound source. In other words, the probability that a sound source is present near an area where the intersections of sound source direction estimation vectors are densely concentrated is high. In this case, when the density of intersections is high, the distance between the center of gravity and the intersections is small, and the value of the sum dj of the reciprocals of the intersection distances is large. By utilizing this property, it is possible to determine which cluster's center of gravity, i.e., the tentative sound source estimation position, is the actual sound source estimation position by comparing the values ​​of dj calculated for each cluster. FIG. 7 illustrates an example of a comparison of the sums dj of the reciprocals of the intersection distances calculated for multiple clusters. In the example of FIG. 7, when the dj calculated for clusters 1 to 12 are compared, the dj of clusters 6, 10, and 12 are clearly larger than those of the other clusters. Therefore, it can be determined that the actual sound source estimation positions are estimated to be three sound sources corresponding to the center of gravity positions of clusters 8, 10, and 12. As a judgment criterion based on the value of dj, for example, as shown in Fig. 7, a method is conceivable in which the value of dj for the cluster showing the maximum value of dj is normalized to 1.0, and the judgment threshold for the cluster including the sound source is set to, for example, 0.7. This judgment threshold can be determined experimentally through sound source position estimation using an actual sound source. Alternatively, without setting such a judgment threshold, clustering may be performed by setting the number of clusters to 2 for the sum dj of the reciprocals of the intersection distances, and it may be inferred that the sound source position is estimated to be in the cluster group with a larger average value of dj.

[0057] In step S21, the sound power level calculation unit 114 calculates the sound power level of each sound source based on the estimated sound source position identified by the sound source position estimation unit 113 and the sound pressure level measured by each microphone array 200.

[0058] Although methods for calculating the acoustic power level of a sound source are well known, a simple example will be provided below for clarity. Consider the case where two sound sources S1 and S2 are collected by three microphone arrays A1, A2, and A3. In this case, the sound collected by one microphone array includes contributions from each of the sound sources S1 and S2 as well as noise. The contributions from the sound sources S1 and S2 are the power of the sound sources that are geometrically attenuated (distance attenuation) according to distance. Let Pi (i = 1 to 3) be the power of the sound wave incident on any microphone included in the microphone array (e.g., the microphone at the center of the array), Esj (j = 1, 2), and rsi-Aj be the distance between the sound source and the microphone array. Then, for example, the power P1 incident on any microphone in the microphone array A1 can be expressed by the following equation: ε is the noise component.

number

[0059] Expanding the above equation to a three-microphone array, we obtain the following equation:

number

[0060] By generalizing the above equation, we obtain P=AE, and on both sides we add the generalized inverse matrix A of A. + Multiplying by E=A + The sound source power and noise power are estimated as P. Note that the above generalized formula allows the sound source power to be estimated for any number of sound sources and any number of microphone arrays.

[0061] In step S22, the calculation result output unit 115 outputs the sound source positions estimated by the sound source position estimation unit 113 and the sound power of each sound source calculated by the sound power level calculation unit 114 through an appropriate output device such as a display or printer of the output unit 140. With this data output, the data processing by the environmental sound monitoring device 100 illustrated in Fig. 5 is completed. By repeatedly executing this data processing flow at predetermined time intervals, it is possible to continuously monitor the status of environmental sound generation at the target construction site. This is expected to have the effect of providing basic data for analyzing the relationship between information on noise around the site and work at the construction site, etc.

[0062] <Output of calculation results> An example of the output of the calculation results will now be described with reference to an example. Fig. 8 shows an example of a graphic display of the calculation results. Fig. 8 shows a situation in which three sets of microphone arrays 200 are installed on an irregular hexagonal site F to search for sound sources, and three sound sources SS1 to SS3 are estimated to be present at the positions shown. Fig. 8 is merely an example and does not represent actual monitoring results or the like.

[0063] FIG. 8 uses a contour map format, showing sound source locations and iso-sound pressure lines, obtained by connecting locations with equal sound pressure levels, with dashed lines. This allows users to intuitively visually grasp the sound pressure distribution within site F and at the boundary with adjacent land. The figures showing sound sources SS1 to SS3 may vary in color or brightness depending on the intensity of the sound power. Furthermore, continuously and repeatedly acquired calculation results may be stored in the memory unit 120, and the sound source locations, the sound power of each sound source, and temporal changes in the iso-sound pressure lines may be displayed as moving images. Alternatively, instead of using a graphic display, the calculated position coordinates of the estimated sound source locations, sound power values, etc. may be displayed numerically. Although not shown in FIG. 8 for simplicity, the site F may be displayed superimposed on a surrounding map. In this case, the range of the noise zone NZ around site F, which has been input as information, may be displayed.

[0064] The above-described embodiment makes it possible to grasp the positions of multiple sound sources occurring within a detection area such as a construction site. Furthermore, by grasping the positions of the sound sources, it becomes possible to estimate the power level of each sound source by calculation based on the sound pressure level recorded by each microphone array 200 and the positions of the microphone array 200 and the sound source. As a result, by grasping the sound sources with a large influence, it becomes possible to take measures against sound sources that have a large impact on the environment around the site.

[0065] Furthermore, by understanding the position of the sound source and the sound power level, it is possible to perform a sound propagation simulation inside and outside the monitored area based on this information. If the monitored area is the site of the site, the impact of the sound source on the surrounding area of ​​the site can be simulated by inputting the buildings around the site into the environmental sound monitoring device 100 in advance. When the position of the sound source and the sound power level change, this type of simulation can be performed by estimating the sound source position and sound power level and performing the simulation again, thereby creating an environmental sound map of the surrounding area of ​​the site at that time. By periodically performing sound source position detection and sound simulation, it is possible to know the time series changes in the environmental sound map.

[0066] As described above, the environmental sound monitoring device 100 according to this embodiment includes a sound source position estimation unit 113 that calculates a sound source estimation direction vector for each of a plurality of microphone arrays 200 based on audio signals from the plurality of microphone arrays 200, each of which has a plurality of microphones 210, calculates planar position coordinates of all intersections of the sound source estimation direction vectors obtained for each of the microphone arrays 200, classifies the intersections into a predetermined number of clusters, calculates the center of gravity of the intersections to which each cluster belongs, and determines an estimated sound source position for each of the clusters based on an index indicating the density of the intersections relative to the center of gravity.

[0067] In this way, the position of the source of environmental noise at a construction site or the like can be determined with high accuracy by simple calculation processing.

[0068] The system may further include an acoustic power level calculation unit 114 that calculates an acoustic power level of each of the sound sources based on the determined estimated sound source positions and the sound pressure levels measured by each of the microphone arrays 200 .

[0069] In this way, the acoustic power of the sound source that generates environmental noise at a construction site, etc. can be accurately determined using simple calculation processing, and can be used to analyze the correlation with surrounding noise, etc.

[0070] For each of the clusters, the sum of the reciprocals of the distances between the center of gravity and each of the intersections is calculated, and the center of gravity of the cluster that is greater than a predetermined threshold can be identified as the estimated position of the sound source.

[0071] In this way, the computational load for estimating the sound source position can be reduced.

[0072] For each cluster, the sum of the reciprocals of the distances between the center of gravity and each of the intersections is calculated, and the values ​​of the sums of the reciprocals of the distances between the center of gravity and each of the intersections obtained for each cluster are clustered into two clusters. Of the two obtained clusters, the center of gravity of the cluster with the larger average value can be identified as the sound source estimation position.

[0073] In this way, the computational load for estimating the sound source position can be reduced.

[0074] The above-described series of processes can be executed by hardware or software. In other words, the functional configuration of FIG. 4 is merely an example and is not particularly limited. That is, it is sufficient for the environmental sound monitoring device 100 to have the functionality to execute the above-described series of processes as a whole, and the functional blocks used to realize these functions are not particularly limited to the example of FIG. 4. Furthermore, one functional block may be configured as a single piece of hardware, a single piece of software, or a combination thereof. The functional configuration of this embodiment is realized by a processor that executes arithmetic processing. Processors that can be used in this embodiment include those configured as various processing units alone, such as single processors, multiprocessors, and multicore processors, as well as those that combine these various processing units with processing circuits, such as ASICs (Application Specific Integrated Circuits) and FPGAs (Field-Programmable Gate Arrays).

[0075] When a series of processes is executed by software, the programs that make up the software are installed into a computer or the like from a network or a recording medium. The computer may be a computer built into dedicated hardware. Alternatively, the computer may be a computer that can execute various functions by installing various programs, such as a general-purpose personal computer.

[0076] The recording medium containing such a program may be a removable medium such as a USB memory that is distributed separately from the device main body in order to provide the program to the user, or may be a recording medium that is provided to the user in a state where it is pre-installed in the device main body. Removable media may be, for example, a magnetic disk (including a floppy disk), an optical disk, or a magneto-optical disk. Optical disks may be, for example, CD-ROMs (Compact Disk-Read Only Memory), DVDs (Digital Versatile Disks), Blu-ray (registered trademark) Discs, etc. Magneto-optical disks may be, for example, MDs (Mini-Disks), etc. Furthermore, recording media that are provided to the user in a state where they are pre-installed in the device main body may be, for example, a ROM on which the program is recorded, a hard disk included in the storage unit 120, a semiconductor memory, etc.

[0077] In this specification, the steps of describing a program to be recorded on a recording medium include not only processes that are performed chronologically in accordance with the order, but also processes that are not necessarily performed chronologically but are performed in parallel or individually.

[0078] Although several embodiments of the present invention have been described above, these embodiments are merely illustrative and do not limit the technical scope of the present invention. The present invention can take various other embodiments, and it is also possible to combine the configurations of the above embodiments and modifications. Furthermore, various modifications such as omissions and substitutions can be made without departing from the spirit of the present invention. These embodiments and modifications thereof are included within the scope and spirit of the invention described in this specification, etc., and are included in the invention described in the claims and their equivalents. [Explanation of symbols]

[0079] 1. Environmental sound monitoring system 100 Environmental sound monitoring device 200 microphone array (sound collection unit) 210 Microphone (sound collection means) 110 Processing section 112 Arrival direction calculation unit 113 Sound source position estimation section 114 Acoustic power level calculation unit 115 Calculation result output section SS sound source

Claims

1. calculating a sound source estimation direction vector pointing to the position of the sound source for each of a plurality of sound collection units based on sound signals from the plurality of sound collection units, each of which includes a plurality of sound collection means; Calculating the plane position coordinates of all intersections of the sound source estimation direction vectors obtained for each of the sound collection units; classifying the plurality of intersections into a predetermined number of clusters, and calculating the center of gravity of the intersections belonging to each cluster; a sound source position estimating unit that determines an estimated sound source position based on an index indicating the density of the intersections with respect to the center of gravity for each of the clusters; Equipped with environmental sound monitoring equipment.

2. The environmental sound monitoring device according to claim 1, further comprising a sound source analysis unit that calculates the acoustic power level of each of the sound sources based on the determined sound source estimated position, the position of each of the sound collection units, and the sound pressure level measured by each of the sound collection units.

3. 2. The environmental sound monitoring device according to claim 1, wherein the sum of the reciprocals of the distances between the center of gravity and each of the intersections is calculated for each of the clusters, and the center of gravity of the cluster for which the sum is greater than a predetermined threshold is identified as the estimated position of the sound source.

4. 2. The environmental sound monitoring device according to claim 1, wherein the sum of the reciprocals of the distances between the center of gravity and each of the intersections is calculated for each of the clusters, the sum is clustered into two clusters, and the center of gravity of the cluster with the larger average value of the two clusters obtained is identified as the estimated position of the sound source.

5. The computer calculating a sound source estimation direction vector pointing to the position of the sound source for each of a plurality of sound collection units based on sound signals from the plurality of sound collection units, each of which includes a plurality of sound collection means; Calculating the plane position coordinates of all intersections of the sound source estimation direction vectors obtained for each of the sound collection units; classifying the plurality of intersections into a predetermined number of clusters, and calculating the center of gravity of the intersections belonging to each cluster; determining an estimated position of a sound source for each of the clusters based on an index indicating the density of the intersections relative to the center of gravity; Environmental sound monitoring method.

6. On the computer, calculating a sound source estimation direction vector pointing to the position of the sound source for each of a plurality of sound collection units based on sound signals from the plurality of sound collection units, each of which includes a plurality of sound collection means; Calculating the plane position coordinates of all intersections of the sound source estimation direction vectors obtained for each of the sound collection units; classifying the plurality of intersections into a predetermined number of clusters, and calculating the center of gravity of the intersections belonging to each cluster; executing a process of determining an estimated position of a sound source based on an index indicating the density of the intersections with respect to the center of gravity for each of the clusters; program.

Images