Sound source display system and sound source display method

Abstract

Providing technology for appropriately indicating sources of noise at construction sites. [Solution] The sound source display system includes a storage unit that stores acoustic signals related to various sounds that may occur at a construction site, associating them with the type of work or tools that can generate those sounds and images representing them; a sound collection unit that collects sounds and converts them into signals; an analysis unit that analyzes the input signals from the sound collection unit, identifies the type of work or tools corresponding to each sound included in the input signal, and identifies the noise level of each sound; and a display unit 40 that instantly visualizes information about noise generated at the construction site according to the analysis results by the analysis unit. The display unit 40 is installed on the outer surface TF of the temporary enclosure at the construction site, and displays a noise information screen 41 when the sound level exceeds a predetermined threshold, and updates the screen display periodically. The noise information screen 41 displays an illustration 42 showing the type of work or tools identified as the noise source and a noise level display 43 in parallel, showing a detailed breakdown of the noise.

Description

[Technical Field]

[0001] The present invention relates to a system and a method for displaying information related to noise sources at a construction site. [Background Art]

[0002] At a construction site, information related to the content of construction work is provided to the public such as neighboring residents and passersby through displays on temporary enclosures and the like. For example, Patent Document 1 discloses, as a configuration of a construction signboard installed when civil engineering work is performed on a road, that the type of civil engineering work (e.g., "Gas Work in Progress"), construction period, constructor, client company and other information are displayed on a display unit. [Prior Art Documents] [Patent Documents]

[0003] [Patent Document 1] Jitsu-Kohei No. 06-017848 [Summary of the Invention] [Problems to be Solved by the Invention]

[0004] According to the above-described prior art, it is considered that minimum information related to construction work can be provided to the public. However, the display according to this prior art lacks real-time performance, and there is a possibility that it is left unchanged for a while even when the work type changes. Furthermore, only the phrase "Gas Work in Progress" makes it difficult to imagine what kind of work specifically is generating the noise.

[0005] At a construction site, a plurality of work types often proceed simultaneously. Different tools and construction machinery are used according to the work type, generating mixed noise from each of the operations. However, when the provided information is insufficient or does not match the current situation, it is difficult to imagine the noise source, which may increase the public's discomfort toward construction noise.

[0006] Therefore, the present invention aims to provide a technology for appropriately indicating the source of noise at construction sites. [Means for solving the problem]

[0007] To solve the above problems, the present invention employs the following sound source display system and method implemented by this system. Note that the following statements in parentheses are merely examples, and the present invention is not limited thereto.

[0008] In other words, the sound source display system of the present invention includes a display unit that displays information about the type of work or tool that was the source of the sound when the loudness of the sound generated by construction work being carried out at a construction site exceeds a predetermined threshold.

[0009] According to this type of sound source display system, when a sound (=noise) exceeding a predetermined threshold occurs at a construction site, it is possible to appropriately display information about the type of work or tool that caused the sound.

[0010] Preferably, in the sound source display system according to the above embodiment, the system further comprises: a storage unit that stores acoustic signals relating to multiple sounds that may occur at a construction site, associated with the type of work or tools that can generate those sounds, along with images indicating the type of work or tools; a sound collection unit that collects ambient sounds generated at the construction site; and an analysis unit that, when the magnitude of the ambient sound exceeds a predetermined threshold, analyzes the ambient sound and identifies the type of work or tools corresponding to each individual sound based on the similarity between the individual signals corresponding to each sound contained in the ambient sound and the acoustic signals stored in the storage unit, and the display unit displays images indicating the type of work or tools identified by the analysis unit.

[0011] According to this type of sound source display system, when noise occurs at a construction site, the sound is analyzed, the type of work or tool corresponding to each sound contained in the sound is identified, and an image showing the identified type of work or tool is displayed. This allows the public to be informed of what type of work or tool is causing the noise (what kind of construction is currently being carried out), and it can reduce the discomfort caused by the noise compared to displaying only text information.

[0012] More preferably, in the sound source display system of any of the above embodiments, when the analysis unit identifies multiple types of work or tools, the display unit displays an image representing each type of work or tool and the noise level generated by that type of work or tool in parallel.

[0013] According to this embodiment of the sound source display system, when multiple types of work or tools are identified as noise sources, images representing them and their noise levels are displayed side by side, making it possible to provide detailed information about the breakdown of the noise.

[0014] More preferably, in the sound source display system of any of the above embodiments, the analysis unit narrows down the information in the storage unit based on process information including construction period, type of work, method, and tools, to the type of work or method of work or tools to be used that are scheduled to be carried out on the day or within a predetermined period including the days before and after, and then identifies the type of work or tools corresponding to each sound.

[0015] According to this type of sound source display system, when identifying the type of work or tool, the information stored in the storage unit is narrowed down based on the process information. Therefore, even if there is a discrepancy between the planned process and the actual work, it is possible to handle not only the type of work on the day but also the types of work on the preceding and following dates, thereby enabling accurate identification of the type of work that caused the noise and, consequently, the display of information related to the noise.

[0016] Furthermore, preferably, in any of the above-described embodiments of the sound source display system, the display unit displays an illustration showing the type of work or tool corresponding to each sound.

[0017] According to the sound source display system of this aspect, an illustration showing the type of work or tool identified as the noise source is displayed. Therefore, the information related to noise is more user-friendly than when displayed only by text information or photographs, etc., thus softening the public's impression of noise and further reducing discomfort.

[0018] More preferably, in the sound source display system according to any of the above aspects, the display unit is further capable of displaying at least a process schedule related to construction at a construction site or a scheduled completion photograph.

[0019] According to the sound source display system of this aspect, when no noise is generated (when no sound of a magnitude that can be called noise is generated) or when the noise source cannot be identified, information such as the process schedule and scheduled completion photograph can be displayed, thereby facilitating public access to information related to the construction site. [Advantageous Effects of Invention]

[0020] As described above, according to the present invention, appropriate display related to a noise generation source at a construction site can be performed. [Brief Description of Drawings]

[0021] [Figure 1] It is a block diagram showing the configuration of a sound source display system 1 according to an embodiment. [Figure 2] It is a diagram showing an example of the structure of a search area 11 that forms a part of an accumulation unit 10. [Figure 3] It is a flowchart showing an example of the procedure of analysis processing. [Figure 4] It is a schematic diagram showing the flow of analysis processing. [Figure 5] It is a diagram for explaining a display unit 40. [Mode for Carrying Out the Invention]

[0022] The embodiments of the present invention will be described below with reference to the drawings. The following embodiments are preferred examples of sound source display systems, and the present invention is not limited to these examples.

[0023] [Configuration of the sound source display system] Figure 1 is a block diagram showing the configuration of a sound source display system 1 according to one embodiment.

[0024] The sound source display system 1 is broadly comprised of a storage unit 10, a sound collection unit 20, an analysis unit 30, and a display unit 40. The storage unit 10 is a storage system such as a database, and is divided into four areas according to its purpose. The search area 11 stores acoustic signals related to various sounds that may occur at a construction site, associating them with the type of work or tools that can generate those sounds.

[0025] Figure 2 shows an example of the structure of the search area 11. Search area 11 contains master tables for acoustic signals related to various sounds that may occur at a construction site, as well as master tables related to work types, construction methods, and tools. Figure 2 is an example of a view obtained by linking these tables. As shown in Figure 2, there are various types of work, such as rebar work, steel frame work, concrete work, and formwork work, and the types of tools that can be used in each work type are associated with these. Some tools, such as tools A and B in the figure, are used in multiple work types. Acoustic signals related to the sounds emitted from each work type or tool are associated with the work type or tool. Although not shown in the figure, images (still images, videos) such as illustrations that represent tools and work types are also associated with and stored.

[0026] It should be noted that the structure described above is merely an example, and the structure of the search area 11 is not limited to this and can be modified as appropriate. For the sake of explanation, the specific names of the tools are omitted here, but the actual table contains their specific names. In addition, in the illustrated example, each acoustic signal is associated with a specific type of work or tool, but the search area 11 also stores acoustic signals related to sounds that cannot be associated with a specific type of work, such as human shouts and sounds generated during transportation.

[0027] Returning to Figure 1, we will continue to explain the configuration of the sound source display system 1. The process area 12 is an area for storing information related to the processes carried out at the construction site. Information such as the construction period, type of work, construction method, and tools is entered into the process area 12 by construction personnel (e.g., site managers) via a simple selection-based input form using information terminals such as smartphones and tablet PCs.

[0028] The synthesis area 13 is equipped with a function for automatically synthesizing acoustic signals. The synthesis area 13 stores acoustic signals synthesized using this function, specifically, various synthesized acoustic signals obtained by superimposing multiple acoustic signals stored in the search area 11 in an arbitrary composition ratio (volume balance). The sound collection area 14 is an area for storing signals corresponding to sounds from the construction site obtained by the sound collection unit 20, which will be described later, as needed.

[0029] The sound-collecting unit 20 is a microphone or a sound level meter that collects sounds generated at the construction site and converts them into electrical signals. The sound-collecting unit 20 is installed inside the temporary enclosure at the construction site.

[0030] The analysis unit 30 analyzes the electrical signals (hereinafter referred to as "input signals") input from the sound collection unit 20, identifies the type of work or tool corresponding to each sound contained in the input signals, and identifies the noise level of each sound based on the noise level measured by the sound collection unit 20. The analysis unit 30 has a neural network model 31 that has been pre-trained using acoustic signals stored in the search area 11 and synthesis area 13 of the storage unit 10 as training data, and uses this to analyze the input signals. The details of the analysis processing performed by the analysis unit 30 will be described in detail later with reference to another drawing.

[0031] The display unit 40 is a display device (for example, digital signage) installed on the outside of the temporary enclosure. When the loudness of the sound generated at the construction site exceeds a sound pressure level (a predetermined threshold) that can be considered noise, it visualizes information about the noise generated at the construction site in real time, according to the analysis results of the analysis unit 30. Specifically, the display unit 40 displays images of the type of work or tools identified as noise sources by the analysis unit 30, and the noise levels of the sounds generated by each type of work or tool, in parallel. A specific example of the display by the display unit 40 will be described later using another drawing.

[0032] The storage unit 10 and the analysis unit 30 are implemented in a computer. They may be implemented in the same computer, or they may be implemented in separate computers and connected via a network, provided that the environment does not cause delays that would affect processing.

[0033] [Analysis Processing] Figure 3 is a flowchart showing an example of the analysis process. The analysis process is performed by the analysis unit 30 at regular intervals (for example, every 10 seconds) on the input signal for the most recent predetermined time. The procedure will be explained below according to the example.

[0034] Steps S1-S3: The analysis unit 30 receives an input signal corresponding to the sound collected by the sound collection unit 20 (Step S1), and converts the input signal for the most recent predetermined time into an amplitude spectrogram using a Short-Time Fourier Transform (hereinafter abbreviated as "STFT") (Step S2). The amplitude spectrogram represents the power of the signal from the result of calculating the frequency spectrum by passing the input signal through a window function. By performing STFT, the change in frequency over time can be analyzed quickly and in detail while maintaining frequency resolution. Subsequently, the analysis unit 30 inputs the amplitude spectrogram into the neural network model 31 (Step S3).

[0035] The neural network model 31 uses a Deep U-Net model and is pre-trained to output mask data necessary for extracting constituent sounds from the amplitude spectrogram of the input signal, using acoustic signals (single acoustic signals) related to individual sounds that may occur at a construction site, which are stored in the search region 11, and synthesized acoustic signals stored in the synthesis region 13, which is formed by superimposing multiple single acoustic signals in various composition ratios, as training data.

[0036] Step S4: The neural network model 31 calculates the masks necessary to extract individual sounds from the input signal based on the amplitude spectrogram input in step S3, and outputs mask data represented as a matrix with frequencies in the rows and time frames in the columns. The mask calculation is performed for each type of sound contained in the input signal (for example, if the input signal contains three overlapping sounds, the calculation is performed for three different types of sounds).

[0037] Step S5: The analysis unit 30 sequentially multiplies the mask data obtained in step S4 with the input signal and its phase spectrogram, and performs an inverse short-time Fourier transform (hereinafter abbreviated as "ISTFT"). This extracts the signals corresponding to each individual sound contained in the input signal (hereinafter referred to as "individual signals").

[0038] Step S6: The analysis unit 30 further identifies the noise level of each sound based on the noise level measured by the sound collection unit 20 with respect to the individual signals extracted in step S5 above.

[0039] Step S7: The analysis unit 30 then narrows down the data stored in the search area 11 of the storage unit 10 based on the information stored in the process area 12, using the type of work, method, or tool to be used that is scheduled to be carried out on the day or within a predetermined period including the days before and after, and then searches for acoustic signals that have a high degree of match with the individual signals obtained in step S5 above.

[0040] Step S8: The analysis unit 30 identifies the tools or work types associated with the acoustic signals that had a high degree of matching with the individual signals in the search area 11 as noise sources. Once the above steps are completed, the analysis process will be finished.

[0041] In the above example, the analysis unit 30 performs real-time analysis on input signals from the most recent predetermined time. However, it is also possible to temporarily store the input signals in the sound collection area 14 of the storage unit 10 and then perform retrospective analysis on the data stored in the sound collection area 14 (i.e., input signals obtained in the past).

[0042] Furthermore, in the above example, the sound source is identified for all individual signals extracted from the input signal. Alternatively, individual signals may be partially extracted based on physical indicators such as noise level and frequency, and the sound source of those individual signals may be identified. For example, the sound source may be identified only for individual signals extracted from the input signal whose noise level is above a certain level. This not only makes it possible to identify the type of work or tool that greatly affects the level of discomfort, but also helps to improve the construction methods for such types of work.

[0043] Furthermore, if the analysis unit 30 is unable to extract individual signals from the input signal, or is unable to identify the type of work or tool corresponding to the individual signal extracted from the input signal, that is, if no acoustic signal similar to the individual signal exists in the search area 11, the input signal or individual signal may be newly stored in the search area 11, and this signal may be added to the training data to further train the neural network model 31. Note that for the signals stored in this way, it is necessary to separately manually associate the type of work being performed or the tool being used when the signal was acquired with the image representing it. By using this configuration, the accuracy of sound source identification can be improved.

[0044] Figure 4 schematically illustrates the flow of the analysis process to facilitate understanding of the invention. In the analysis process, the input signal shown in Figure 4(A) is subjected to a series of processes shown in steps S1 to S5 in Figure 3, thereby extracting individual signals corresponding to each sound contained in the input signal, as shown in Figure 4(B). Then, as shown in Figure 4(C), the search area 11 is narrowed down based on information about the process, and an acoustic signal with a high degree of match to the individual signals is searched for.

[0045] The illustrated example shows the result of filtering when "rebar work" and "steel frame work" are scheduled to be carried out on the day or within a specified period including the days before and after, and "tool H" is not included in the tools to be used. The data excluded by filtering is shown in shaded areas. After this filtering is applied to the data in search area 11, acoustic signals with a high degree of match with individual signals are searched from the acoustic signals in search area 11, and the type of work or tool associated with that acoustic signal is identified as the sound source. In the illustrated example, as a result of the search, three acoustic signals with a high degree of match with three individual signals (acoustic signals marked with circles in the figure) are found, and finally, the type of work or tool associated with these is identified as the sound source.

[0046] In this way, by narrowing down the data in the search area 11 based on process information, the accuracy of identifying the sound source can be improved. Furthermore, since the narrowing down is based on process information for a predetermined period including not only the day but also the days before and after, it can also handle cases where there is a discrepancy between the entered process and the actual work due to the progress of the construction being slightly behind schedule or ahead of schedule.

[0047] [Noise-related information] Figure 5 shows the display unit 40. As shown in Figure 5(A), the display unit 40 is installed, for example, on the outer surface of the temporary enclosure TF at the construction site. When the noise level generated at the construction site exceeds a predetermined threshold, as shown in Figure 5(B), the display unit 40 displays the noise information screen 41 and updates the screen display at regular intervals.

[0048] The noise information screen 41 displays the current time, along with illustrations 42 showing the specific types of work or tools identified as noise sources, and a noise level display 43 for those noise sources. This display allows nearby residents, passersby, and other members of the public to be informed in real time what kind of construction is currently being carried out inside the temporary enclosure and how loud each type of work or tool is generating.

[0049] The paper "Manipulation of Environmental Impressions through the Interaction of Vision and Auditory Perception" (by Yuko Masakura et al., VISION, Japanese Society for Vision Science, Vol. 15, No. 3, pp. 117-132, July 2003) discloses the results of an experiment conducted to investigate the interaction of vision and auditory perception that influences noise and impressions of the environment. The experiment found that, compared to cases where no images were presented or images with low harmony with the noise source were presented, presenting images appropriate to the noise source reduced discomfort and improved impressions. Based on these findings, it is considered that when construction noise occurs, presenting visual information appropriate to the construction noise will reduce discomfort and improve impressions.

[0050] Therefore, in this embodiment, the sound sources of construction noise are identified in real time, and illustrations showing the type of work or tools identified as each sound source are displayed along with their respective noise levels. By displaying such noise information, the discomfort caused by construction noise can be reduced compared to when unrelated information is displayed. Furthermore, displaying the type of work or tools with illustrations is more user-friendly than displaying only text information, thus further reducing discomfort from construction noise and fostering an understanding of the construction work. In addition, the illustrations can soften the impression, and an improvement in the streetscape can be expected.

[0051] Furthermore, if the noise level at the construction site is below a predetermined threshold, it means that no noise of a level that could be considered noise has been generated. In such cases, the display unit 40 will display a construction schedule, a photo of the planned completion, or other information using functions attached to the display unit 40 (for example, a content display function installed on a digital signage system) instead of the noise information screen 41 described above. The same display will also be made if the noise source cannot be identified by the analysis unit 30. By providing such displays, the public will be able to access information about the construction site more easily.

[0052] As explained above, the following effects can be obtained using the sound source display system 1 described above.

[0053] (1) An input signal containing one or more sounds occurring simultaneously at a construction site is analyzed, and the individual signals for each sound source included in the input signal are identified as belonging to which type of work or tool. This makes it possible to identify the type of work or tool that caused the noise and to understand the breakdown of the noise.

[0054] (2) Prior to searching the storage unit 10 (search area 11) for acoustic signals similar to the individual signals extracted from the input signals, the data is narrowed down based on the information in the process area 12 by the type of work or method of work or the tools to be used that are scheduled to be carried out on the day or within a predetermined period including the days before and after. Therefore, even if there is a discrepancy between the input (scheduled) process and the actual work, the noise source can be identified and displayed with high accuracy.

[0055] (3) The search area 11 stores not only acoustic signals related to various sounds that may occur at a construction site, but also acoustic signals related to sounds that cannot be associated with a specific type of work, such as shouts from people and sounds generated during transportation. This makes it possible to identify such sound sources as well, thereby improving the accuracy of identifying and displaying noise sources.

[0056] (4) During the analysis, individual signals can be partially extracted based on physical indicators such as noise level and frequency, and the sound sources of these individual signals can be identified and displayed, thereby enabling the identification and display of construction types that significantly affect discomfort, and helping to improve the construction methods for those types of work.

[0057] (5) The neural network model 31 used in the analysis process learns various patterns related to sound overlap by using various single acoustic signals accumulated in the search region 11 and various synthesized acoustic signals accumulated in the synthesis region 13, which is formed by superimposing multiple single acoustic signals in an arbitrary composition ratio, as training data. Therefore, it can handle various types of noise with different volume balances when multiple sounds occur simultaneously.

[0058] (6) If, during analysis, it is not possible to extract individual signals from the input signal, or if it is not possible to identify the type of work or tool corresponding to the individual signals extracted from the input signal, this signal can be newly stored in the search area 11, and by adding this signal to the training data and further training the neural network model 31, the accuracy of identifying and displaying noise sources can be improved.

[0059] (7) The input signal is analyzed and the noise information as a result of the analysis is displayed in real time, and the type of work or tool that was the source of the noise is displayed, so that the public's discomfort with construction noise can be reduced compared to when information unrelated to the noise source is displayed.

[0060] (8) In the search area 11, various acoustic signals are linked to types of work or tools, and images representing the types of work or tools are also linked and stored. As a result, the types of work or tools identified as noise sources through analysis can be displayed as images, which can reduce the discomfort caused by construction noise compared to when only text information is displayed.

[0061] (9) When multiple types of work or tools are identified as noise sources, images representing them and their noise levels are displayed side by side, allowing for a detailed breakdown of construction noise and further reducing discomfort from construction noise.

[0062] (10) Since the type of work or tools that are the source of the noise are displayed in illustrations, it is easier to understand than when only text information or photographs are used, which can soften the public's impression of construction noise and further reduce discomfort.

[0063] The present invention can be implemented in various ways without being limited to the embodiments described above.

[0064] In the embodiment described above, the storage unit 10 has four regions (search region 11, process region 12, synthesis region 13, and sound collection region 14) according to the purpose, but in addition to these regions, a results region may be provided for storing the history of analysis results by the analysis unit 30. This makes it possible to check the noise situation over a specific period in the past.

[0065] In the embodiment described above, information regarding the process is entered into the process area 12 by construction personnel via a simple input form. However, if similar information is entered into another system established for a different purpose, that information may be periodically imported into the process area 12 and used while being synchronized. This prevents construction personnel from having to enter information twice, while effectively utilizing information entered into another system.

[0066] In the embodiment described above, the analysis unit 30 newly stores in the search area 11 any input signals from which it could not extract individual signals, or any individual signals from which it could not identify the type of work or tool, and adds them to the training data, thereby enabling the retraining of the neural network model 31. This retraining may be performed sequentially each time new data is accumulated, or it may be performed all at once after a certain amount of new data has been accumulated.

[0067] Furthermore, all other examples shown with illustrations in the embodiments are merely preferred examples and can be modified as appropriate when implementing the present invention. [Explanation of Symbols]

[0068] 1. Sound Source Display System 10 Storage section 20 Sound collection section 30 Analysis Department 40 Display section 41 Noise Information Screen 42 Illustrations 43 Noise level display

Claims

1. A display unit that, when the noise level generated by construction work being carried out at a construction site exceeds a predetermined threshold, displays information about the type of work or tool that caused the noise. A sound source display system equipped with this feature.

2. In the sound source display system according to claim 1, A storage unit stores acoustic signals related to multiple sounds that may occur at a construction site, associated with the type of work or tool that can generate those sounds, along with images representing the said type of work or tool. A sound collection unit that collects ambient noise generated at the construction site, When the magnitude of the ambient sound exceeds the predetermined threshold, the analysis unit analyzes the ambient sound and identifies the type of work or tool corresponding to each sound based on the similarity between the individual signals corresponding to each sound contained in the ambient sound and the acoustic signals stored in the storage unit. Furthermore, The aforementioned display unit is A sound source display system characterized by displaying an image showing the type of work or tool identified by the analysis unit.

3. In the sound source display system according to claim 2, The aforementioned display unit is A sound source display system characterized in that, when the analysis unit identifies multiple types of work or tools, it displays an image representing each type of work or tool and the noise level generated by that type of work or tool in parallel.

4. In the sound source display system according to claim 3, The aforementioned analysis unit, A sound source display system characterized by narrowing down the information in the storage unit based on process information including construction period, type of work, method, and tools, to the type of work or method or tools to be used that are scheduled to be carried out on the day or within a predetermined period including the days before and after, and then identifying the type of work or tools corresponding to each individual sound.

5. In the sound source display system according to claim 4, The aforementioned display unit is A sound source display system characterized by displaying an illustration showing the type of work or tool corresponding to each of the aforementioned sounds.

6. In the sound source display system according to any one of claims 1 to 5, The aforementioned display unit is A sound source display system characterized by being able to further display at least a construction schedule or a photograph of the planned completion date related to the construction work at the construction site.

7. A display process that, when the noise level generated by construction work being carried out at a construction site exceeds a predetermined threshold, displays information about the type of work or tool that was the source of the noise. A method for displaying sound sources, including the sound source itself.

Images