Fire detection devices

Abstract

To provide a fire detection device which can detect occurrence of fire and grasp a situation in a monitoring area after detecting the fire on the basis of acoustic data to be collected in the monitoring area.SOLUTION: A fire detection device comprises: a sound collection device which collects sound which is made in a monitoring area as acoustic data; and a fire detection unit which executes fire detection processing of detecting that fire occurs in the monitoring area on the basis of the acoustic data collected by the sound collection device. The fire detection unit executes the fire detection processing by using a high sensitivity level set to sensitivity higher than an initial sensitivity level and grasps a situation in the monitoring area when it is detected that the fire occurs by executing the fire detection processing by using the preset initial sensitivity level.SELECTED DRAWING: Figure 1

Description

[Technical Field]

[0001] The present disclosure relates to a fire detection device that detects the occurrence of a fire in a monitored area based on acoustic data. [Background technology]

[0002] Various types of fire detectors are used to detect the occurrence of a fire in a monitored area, including differential spot detectors, constant temperature spot detectors, and photoelectric spot detectors.

[0003] There are also fire detection devices that detect the sounds that are generated when something burns during a fire (see, for example, Non-Patent Document 1). Non-Patent Document 1 discloses a method for conducting combustion experiments to capture the characteristics of sounds that occur during combustion, and for distinguishing between fire and normal conditions based on frequency analysis of the sounds, thereby determining whether a fire has occurred.

[0004] In particular, Non-Patent Document 1 describes that it is appropriate to specify a frequency range and identify the presence or absence of a combustion phenomenon based on the time change in the integral value of the overall power (level) rather than tracking a specific frequency and level. [Prior art documents] [Non-patent literature]

[0005] [Non-Patent Document 1] "Frequency Analysis of Combustion Sound (2nd Report)," Fire and Disaster Management Research Institute Bulletin No. 31 (1994) Summary of the Invention [Problem to be solved by the invention]

[0006] However, the conventional techniques have the following problems. When conducting fire monitoring, it is important not only to determine whether a fire has occurred or not, but also to grasp the situation within the monitoring area after a fire is detected, such as the extent of the fire's spread, whether the fire has been extinguished, and the status of people who are unable to escape.

[0007] In Non-Patent Document 1, the idea of ​​time change in the integral value of the power spectrum of sound is introduced as a measure to more reliably separate background noise from combustion sounds, and whether or not a fire has occurred is determined based on sounds picked up within the monitoring area. However, Non-Patent Document 1 does not take into consideration grasping the situation after a fire has been detected.

[0008] The present disclosure has been made to solve the above-mentioned problems, and aims to provide a fire detection device that can detect the occurrence of a fire based on acoustic data collected in a monitored area, and can grasp the situation within the monitored area after detecting a fire. [Means for solving the problem]

[0009] A fire detection device according to the present disclosure includes a sound collection device that collects sounds generated within a monitoring area as acoustic data, and a flame detection unit that executes a flame detection process to detect the occurrence of a flame in the monitoring area based on the acoustic data collected by the sound collection device, and when the flame detection unit detects the occurrence of a flame by executing the flame detection process using a preset initial sensitivity level, it executes the flame detection process using a high sensitivity level that is set to a sensitivity higher than the initial sensitivity level. Continue Run, After flame detection Grasp the situation within the monitoring area This makes it possible to detect whether a fire has not been extinguished as one way of understanding the situation. It is something.

[0010] Furthermore, the fire detection device according to the present disclosure is a fire detection device including a sound collection device that collects sounds generated within a monitoring area as acoustic data, and a flame detection unit that executes a flame detection process to detect the occurrence of a flame in the monitoring area based on the acoustic data collected by the sound collection device, wherein the sound collection devices are installed as a plurality of sound collection devices within the monitoring area, and the flame detection unit executes the flame detection process based on the acoustic data collected by each of the plurality of sound collection devices, and when the occurrence of a flame is detected by at least one of the plurality of sound collection devices, Includes one sound pickup device For multiple sound collection devices, a high sensitivity level set higher than the initial sensitivity level set in advance is used to perform flame detection processing. Continue Run, After flame detection Grasp the situation within the monitoring area This makes it possible to detect whether a fire has not been extinguished as one way of understanding the situation. It is something. In addition, the fire detection device according to the present disclosure is a fire detection device that includes a sound collection device that collects sounds generated within a monitored area as acoustic data, and a flame detection unit that executes a flame detection process to detect the occurrence of a flame in the monitored area based on the acoustic data collected by the sound collection device. When the flame detection unit detects the occurrence of a flame by executing the flame detection process using a preset initial sensitivity level, it executes the flame detection process using a high sensitivity level that is set to a higher sensitivity than the initial sensitivity level, and grasps the situation within the monitored area. When the occurrence of a flame is detected using the initial sensitivity level, it switches a human detection mode for detecting human voices to an on setting, and determines whether or not a person is present in the monitored area by executing a pre-trained voice recognition process on the acoustic data collected by the sound collection device. [Effects of the Invention]

[0011] According to the present disclosure, a fire detection device can be obtained that can detect the occurrence of a fire based on acoustic data collected in a monitored area and grasp the situation within the monitored area after detecting a fire. [Brief explanation of the drawings]

[0012] [Figure 1] 1 is an explanatory diagram showing the overall configuration of a fire detection device according to a first embodiment of the present disclosure. [Figure 2] FIG. 1 is an explanatory diagram showing an outline of fire detection in Prior Application 1. [Figure 3] FIG. 1 is an explanatory diagram showing an outline of fire detection in Prior Application 2. [Figure 4] FIG. 1 is an explanatory diagram showing an outline of fire detection by a crib in Prior Application 3. [Figure 5] 1 is an explanatory diagram showing a state in which a monitoring area is monitored by a fire detection device according to a first embodiment of the present disclosure. [Figure 6]FIG. 4 is an explanatory diagram showing a specific example of changing the determination threshold from an initial sensitivity level to a high sensitivity level in the flame detection unit according to the first embodiment of the present disclosure. [Figure 7] 3 is a flowchart showing a series of processes related to a fire detection method executed in the fire detection device according to the first embodiment of the present disclosure. DETAILED DESCRIPTION OF THE INVENTION

[0013] Hereinafter, preferred embodiments of the fire detection device of the present disclosure will be described with reference to the drawings. The fire detection device disclosed herein has the technical feature of detecting the occurrence of a fire based on acoustic data collected in a monitored area, and after detecting a fire, changing the sensitivity level to a high sensitivity level higher than the initial value, thereby continuing to monitor the situation within the monitored area after detecting a fire.

[0014] Embodiment 1 1 is an explanatory diagram showing the overall configuration of a fire detection device according to the first embodiment of the present disclosure. The fire detection device according to the first embodiment includes a microphone 10 and a computer 20.

[0015] The microphone 10 is a sound collection device that is installed in the monitoring area of ​​the flame and collects sounds generated within the monitoring area as acoustic data. Note that it is conceivable to install one or more microphones 10 within the monitoring area.

[0016] When determining whether or not there is a fire based on the acoustic data collected by the microphone 10, the focus is on the acoustic data generated by the combustion itself, rather than monitoring the smoke generated as a by-product of the fire, allowing for faster fire detection. Furthermore, the microphone 10 does not need to be installed on the ceiling like a smoke detector, allowing for greater freedom in determining the installation location.

[0017] The computer 20 is a controller that determines whether a flame has occurred in the monitored area by performing arithmetic processing on the acoustic data collected by the microphone 10, and is equipped with an acoustic data analysis processing unit 21 and a flame detection unit 22.

[0018] Techniques for determining whether a flame has occurred based on acoustic data are described in detail in the following three prior applications by the inventor of the present application: (1) Prior application 1 (Patent application 2022-166968) (2) Prior application 2 (Patent application 2022-166969) (3) Prior application 3 (Patent application 2022-166970)

[0019] The fire detection in each of Prior Applications 1 to 3 will be outlined below with reference to FIGS. 2 to 4. FIG.

[0020] (1) Overview of fire detection in Prior Application 1 Prior application 1 extracts features specific to a flame from the results of frequency analysis based on acoustic data collected in a monitored area to determine whether a flame has occurred, suppressing false alarm factors and achieving high-precision flame detection. In particular, Prior application 1 detects fires by focusing on the 1 / f fluctuation characteristics that pass through peak frequencies in the frequency band below standing waves as a feature specific to a flame.

[0021] Figure 2 is an explanatory diagram showing an overview of fire detection in Prior Application 1. Figure 2(A) is a comparative diagram showing frequency spectra for three types of acoustic data: background noise, flame, and ventilation fan. Specifically, Figure 2(A) shows the three frequency spectra as well as 1 / f fluctuation characteristics passing through the peak frequencies for each of the background noise, flame, and ventilation fan in the frequency band from f1 = 2.1 Hz to f2 = 10.0 Hz, which is the frequency band below the standing wave.

[0022] Figure 2(B) shows the average level error for three types of acoustic data: background noise, flame, and ventilation fan in the frequency band below the standing wave. For comparison, Figure 2(B) also shows the average level error for flame in the frequency band from 2.1 Hz to 20 Hz.

[0023] Here, the "average level error" is introduced as an index value that quantitatively specifies the degree of approximation between the frequency spectrum and the 1 / f fluctuation characteristics. As shown in Fig. 2(B), the flame detection unit 22 can detect a flame, distinguishing it from false alarm factors, by determining whether the average level error is equal to or less than the judgment threshold, for example, set to 4 dB.

[0024] The 4 dB set as the determination threshold is an initial value used to detect the occurrence of a flame, and corresponds to the initial sensitivity level in the present disclosure.

[0025] (2) Overview of fire detection in Prior Application 2 Prior Application 2 extracts features specific to a flame from the results of frequency analysis based on acoustic data collected in a monitored area to determine whether a flame has occurred, suppressing false alarm factors and achieving high-precision flame detection. In particular, Prior Application 2 detects fires by focusing on the level transition over time of low-frequency peak frequencies, which are a characteristic of a flame.

[0026] FIG. 3 is an explanatory diagram showing an overview of fire detection in Prior Application 2. FIG. 3(A) shows the time transition of peak frequency levels in the low frequency range for combustion noise and door opening / closing sounds. Specifically, FIG. 3(A) illustrates an example in which time series data of peak frequencies is subjected to a moving average process every second to determine level transitions over 120 seconds. Comparing the two waveforms shown in FIG. 3(A), it can be seen that the door opening / closing sound has greater fluctuations in level transitions than the combustion sound.

[0027] Figure 3(B) shows the result of calculating the level fluctuation amount, which indicates the degree of variation, from the waveform of Figure 3(A). Here, the "level fluctuation amount" corresponds to the value calculated as the standard deviation from the waveform of the level transition, and is introduced as an index value that quantitatively indicates the fluctuation variation of the level transition.

[0028] As shown in Figure 3(B), the flame detection unit 22 can detect a flame, distinguishing it from the sound of a door opening and closing, which is a cause of a false alarm, by setting a judgment threshold of, for example, 8 dB and determining whether the level fluctuation amount is equal to or greater than the judgment threshold.

[0029] The 8 dB set as the determination threshold is an initial value used to detect the occurrence of a flame, and corresponds to the initial sensitivity level in the present disclosure.

[0030] (3) Overview of fire detection in Prior Application 3 Prior Application 3 extracts features specific to flames caused by burning cribs from the results of frequency analysis based on acoustic data collected in a monitored area, determines whether a flame is generated by a crib, and suppresses false alarm factors, thereby achieving high-precision flame detection. Here, "crib" refers to a general term for a specific combustible material that generates a pulsating sound when burning.

[0031] In particular, Prior Application 3 focuses on the transition of sound pressure levels in the high frequency range, which is a characteristic feature of crib flames, and detects crib fires.

[0032] Figure 4 is an explanatory diagram outlining the fire detection by a crib in Prior Application 3. Figure 4(A) shows the difference in the transition of sound pressure levels for the burning sound of heptane, the burning sound of a crib, and the sound of a ventilation fan. Specifically, Figure 4(A) illustrates how the burning sound of a crib can be identified by comparing the number of times the sound pressure level changes from a value below the judgment level to a value above the judgment level, assuming that the judgment level is set to 15 dB.

[0033] Figure 4(B) shows the results of counting the number of occurrences from the waveform of Figure 4(A). The number of occurrences of the crib combustion sound is greater than the number of occurrences of the heptane combustion sound and the number of occurrences of the ventilation fan sound. As shown in Figure 4(B), the flame detection unit 22 can detect the crib flame, distinguishing it from the ventilation fan sound, which is a cause of false alarms, by determining whether the number of occurrences is equal to or greater than the judgment threshold, for example, 10 times.

[0034] The judgment threshold value of 10 times is an initial value used to detect the occurrence of a flame caused by the crib, and corresponds to the initial sensitivity level in the present disclosure.

[0035] Returning to the explanation of Fig. 1 of the present disclosure, the acoustic data analysis processing unit 21 shown in Fig. 1 calculates a quantitative value for detecting a flame as a feature amount for the acoustic data acquired via the microphone 10 using various methods such as those described in the above-mentioned Prior Applications 1 to 3. Furthermore, the flame detection unit 22 shown in Fig. 1 determines whether or not a flame has occurred based on the result of comparing the calculated feature amount with a determination threshold.

[0036] Next, the functions of the fire detection device according to the first embodiment will be described in detail with reference to Fig. 5 to Fig. 7. In Fig. 1, the acoustic data analysis processing unit 21 and the flame detection unit 22 are described as separate, independent components. However, the flame detection unit 22 may also be configured to execute the processing of the acoustic data analysis processing unit 21. Therefore, in the following, a detailed description will be given assuming that a configuration is adopted in which all processing based on acoustic data is executed by the flame detection unit 22.

[0037] 5 is an explanatory diagram showing a state in which a monitoring area is monitored by the fire detection device according to the first embodiment of the present disclosure. The fire detection device shown in FIG. 5 is configured to include three microphones 10(1) to 10(3) and a flame detection unit 22.

[0038] The microphones 10(1) to 10(3) are installed at different positions within the monitoring area and are sound collection devices that collect sounds generated within the monitoring area as acoustic data. The microphones 10(1) to 10(3) can be installed at any suitable location, such as on the ceiling, the wall of the monitoring area, or on a desk within the monitoring area.

[0039] Therefore, it is easy to add a fire detection function based on acoustic data later or temporarily. In other words, with a relatively simple configuration, it is possible to realize flame detection processing in various monitoring areas.

[0040] The flame detection unit 22 is connected to each of the multiple microphones 10(1) to 10(3), and executes a flame detection process to detect the occurrence of a flame in the monitoring area based on the acoustic data collected by each of the multiple microphones 10(1) to 10(3). As this flame detection process, any one of the above-mentioned Prior Application 1 to Prior Application 3, or a combination thereof, can be applied.

[0041] In the initial stage of the flame detection process, the flame detection unit 22 determines whether a flame has occurred using a preset initial sensitivity level. Then, when the flame detection unit 22 detects the occurrence of a flame from acoustic data collected by at least one of the multiple microphones 10 by performing the flame detection process using the initial sensitivity level, the flame detection unit 22 performs the flame detection process using a high sensitivity level set to a higher sensitivity than the preset initial sensitivity level for the multiple microphones 10.

[0042] For example, in the configuration of FIG. 5, when the flame detection unit 22 detects the occurrence of a flame from the acoustic data collected by the microphone 10(1), it switches the sensitivity of all of the microphones 10(1) to 10(3) to a high sensitivity level that is preset as a sensitivity higher than the initial sensitivity level, and continues the flame detection process.

[0043] 6 is an explanatory diagram showing a specific example of changing the determination threshold from the initial sensitivity level to a high sensitivity level in flame detection unit 22 according to embodiment 1 of the present disclosure. In Fig. 6(A), an arrow indicates a state in which the setting is changed to the high sensitivity level when using the fire detection process according to prior application 1 described with reference to Fig. 2. When using the fire detection process according to prior application 1, if the initial sensitivity level is set to 4 dB, the setting is changed to a value greater than 4 dB as the high sensitivity level.

[0044] In Figure 6(B), the arrow indicates the state in which the setting is changed to a high sensitivity level when using the fire detection process according to Prior Application 2 described with reference to Figure 3. When using the fire detection process according to Prior Application 2, if the initial sensitivity level is set to 8 dB, the setting is changed to a value greater than 8 dB as the high sensitivity level.

[0045] In Figure 6(C), the arrow indicates the state in which the setting is changed to a high sensitivity level when using the fire detection process according to Prior Application 3 described with reference to Figure 4. When using the fire detection process according to Prior Application 3, if the initial sensitivity level is set to 10 times, the setting is changed to a value smaller than 10 times as the high sensitivity level.

[0046] In either case, changing the setting to the high sensitivity level increases the sensitivity for detecting flames, but reduces the margin for false detection factors, increasing the risk of falsely detecting background noise, the sound of a ventilation fan, the sound of a door opening and closing, etc. However, because the setting is changed to the high sensitivity level after a flame has already been detected at the initial sensitivity level, there is no risk of falsely detecting a flame as detected, even though the flame was not detected.

[0047] Furthermore, changing the setting to a high sensitivity level has the advantage of being able to detect with higher sensitivity the spread of a fire, the state of the fire being extinguished, etc. For example, in the configuration of Fig. 5, when the flame detection unit 22 detects the outbreak of a fire from the acoustic data collected by the microphone 10(1), it switches the sensitivity level to high for all of the microphones 10(1) to 10(3) and continues the flame detection process.

[0048] By continuing the flame detection process in this manner, when a flame can be detected using the acoustic data from microphone 10(2) or microphone 10(3) at a high sensitivity level, a smaller flame can be detected than when a flame can be detected using the acoustic data from microphone 10(1).

[0049] Furthermore, regarding the acoustic data from microphone 10(1), which initially contributed to detecting the flames, by switching to a high sensitivity level and continuing monitoring after detecting the occurrence of a flame, it is possible to more accurately detect whether the fire has not been extinguished.

[0050] Furthermore, the flame detection unit 22 according to the first embodiment further has a delayed escape detection function that determines, based on acoustic data, whether there is anyone in the monitored area who has failed to escape. This delayed escape detection function will now be described in detail.

[0051] When the flame detection unit 22 detects the occurrence of a flame using the initial sensitivity level, it changes the setting to a high sensitivity level to continue monitoring, and switches the human detection mode for detecting human voices to the on setting.

[0052] Then, when the human detection mode is set to ON, the flame detection unit 22 executes pre-trained voice recognition processing on the acoustic data collected by each of the multiple microphones 10(1) to 10(3), and performs voice detection.

[0053] That is, in order to detect voices uttered by people within the monitored area, the flame detection unit 22 inputs various acoustic data including voices within the monitored area and creates in advance a voice recognition process as a learning model for identifying the voices.

[0054] When the human detection mode is set to on, the flame detection unit 22 can use the voice recognition processing that it has learned in advance to determine that a person who was unable to escape is present in the monitored area if a voice is detected within the monitored area.

[0055] Next, a series of processes executed in the fire detection device according to the first embodiment will be described using a flowchart. Fig. 7 is a flowchart showing a series of processes related to the fire detection method executed in the fire detection device according to the first embodiment of the present disclosure. Note that the following description will be given taking a specific example in which a plurality of microphones 10 are installed in a monitoring area.

[0056] First, in step S701, the flame detection unit 22 performs a process of collecting sound data generated within the monitoring area via a plurality of microphones 10 installed within the monitoring area.

[0057] Next, in step S702, the flame detection unit 22 executes a flame detection process to detect the occurrence of a flame in the monitoring area based on the acoustic data collected by each of the plurality of microphones 10.

[0058] Next, in step S703, the flame detection unit 22 determines whether or not a flame has been detected as a result of executing the flame detection process. If a flame has been detected, the flame detection unit 22 executes the processes from step S704 onward, and if a flame has not been detected, the series of processes ends.

[0059] If the process proceeds to step S704, the flame detection unit 22 changes the judgment thresholds for the multiple microphones 10 from the initial sensitivity level to a high sensitivity level and continues the flame detection process. As a result, once a flame has been detected, the flame detection unit 22 can detect the spread of the fire, the state of the fire being extinguished, and the like with higher sensitivity. In other words, it is possible to continue monitoring the situation within the monitoring area after detecting a fire.

[0060] Next, in step S705, the flame detection unit 22 switches the human detection mode to ON and executes the delayed escape detection process using the pre-trained voice recognition process. That is, the flame detection unit 22 performs the voice recognition process on the acoustic data to determine whether or not the voice of a person left behind in the monitoring area has been picked up.

[0061] Next, in step S706, the flame detection unit 22 determines whether or not a person who was too late to escape has been detected in the monitoring area as a result of executing the voice recognition process. If a person who was too late to escape has been detected, the flame detection unit 22 executes the process of step S707, and if a person who was too late to escape has not been detected, the series of processes ends.

[0062] In step S706, if the flame detection unit 22 executes voice recognition processing and detects a person in the monitoring area who has not been able to escape in time, the camera that operates in conjunction with this result can be activated, making it possible to use the images captured by the camera to grasp the situation in the monitoring area, including whether there is anyone who has not been able to escape in time.

[0063] If the process proceeds to step S707, the flame detection unit 22 executes a notification process in response to the detection of a person who was too late to escape. For example, the flame detection unit 22 can execute a notification process to guide a person who was too late to escape within the monitored area to an emergency exit or an evacuation route. The flame detection unit 22 can also execute a notification process to notify people outside the monitored area or a higher-level device that a person who was too late to escape within the monitored area has been detected, and to urge rescue.

[0064] In the flowchart of Figure 7, a case where full functionality is performed using multiple microphones through steps S701 to S707 is described, but the fire detection device of the present disclosure is not limited to such full functionality, and it is also possible to adopt configurations 1 to 3 that perform the following processing.

[0065] Configuration 1: A configuration in which the functions of steps S701 to S704 are performed, and the delayed escape detection process using the human detection mode is performed as an option. Configuration 2: A configuration in which only one microphone is used and full functionality is achieved through steps S701 to S707. Configuration 3: A configuration in which only one microphone is used, the functions of steps S701 to S704 are performed, and the delayed escape detection process using the human detection mode is performed as an option.

[0066] Either configuration makes it possible to realize a fire detection device that can detect the occurrence of a fire based on acoustic data collected in the monitored area and grasp the situation within the monitored area after detecting a fire.

[0067] As described above, according to embodiment 1, a fire is detected based on acoustic data collected in the monitoring area, and after a fire is detected, the sensitivity level is changed to a high sensitivity level higher than the initial value, so that situation monitoring within the monitoring area can be continued after a fire is detected.

[0068] In particular, the fire detection device according to the first embodiment provides the following effects. Effect 1: The microphones can be installed in any position, allowing for easy retrofitting or temporary addition of fire monitoring functionality. In other words, a relatively simple configuration can be used to achieve fire detection processing and delayed escape detection processing in various monitoring areas.

[0069] Effect 2: Rather than monitoring smoke generated as a by-product of a fire, the system focuses on the acoustic data generated by the combustion itself to detect fires more quickly. Furthermore, by further performing voice recognition processing on the acoustic data, it is possible to accurately determine whether or not a person is present in the monitoring area.

[0070] Effect 3: By combining the fire detection method according to the present disclosure with a fire detection method using other sensors, it is possible to suppress false alarm factors and more accurately detect whether a fire has occurred. In other words, by taking advantage of the benefits of using acoustic data, it is possible to construct a fire detection method that not only detects the occurrence of a fire but also has the function of grasping the situation within the monitoring area after a fire has been detected.

[0071] Effect 4: By further performing voice recognition processing on acoustic data using the fire detection method disclosed herein, if it is determined that a person is present within the monitored area, the camera can be further linked, and the captured images can be used to grasp the situation within the monitored area, including people who were unable to escape in time. [Explanation of symbols]

[0072] 10 microphone (sound collection device), 20 computer, 21 acoustic data analysis processing unit, 22 flame detection unit.

Claims

1. a sound collection device that collects sounds generated within the monitoring area as acoustic data; a flame detection unit that executes a flame detection process to detect the occurrence of a flame in the monitoring area based on the acoustic data collected by the sound collection device; A fire detection device comprising: The flame detection unit is When the occurrence of a flame is detected by executing the flame detection process using a preset initial sensitivity level, the flame detection process is continued using a high sensitivity level set to a higher sensitivity than the initial sensitivity level, and by grasping the situation within the monitoring area after the flame is detected, it is possible to detect that the fire has not been extinguished as one of the grasping of the situation. Fire detection devices.

2. a sound collection device that collects sounds generated within the monitoring area as acoustic data; a flame detection unit that executes a flame detection process to detect the occurrence of a flame in the monitoring area based on the acoustic data collected by the sound collection device; A fire detection device comprising: the sound collection device is installed as a plurality of sound collection devices within the monitoring area, The flame detection unit is performing the flame detection process based on the acoustic data collected by each of the plurality of sound collection devices; When a flame is detected in at least one of the plurality of sound collection devices, the flame detection process is continued for the plurality of sound collection devices including the one sound collection device using a high sensitivity level set to a sensitivity higher than a preset initial sensitivity level, and by grasping the situation in the monitoring area after the flame detection, it is possible to detect that the fire has not been extinguished as one of the grasping of the situation. Fire detection devices.

3. A sound collection device that collects sounds generated within a monitoring area as acoustic data; a flame detection unit that executes a flame detection process to detect the occurrence of a flame in the monitoring area based on the acoustic data collected by the sound collection device; A fire detection device comprising: The flame detection unit is When the occurrence of a flame is detected by executing the flame detection process using a preset initial sensitivity level, the flame detection process is executed using a high sensitivity level set to a sensitivity higher than the initial sensitivity level, and the situation within the monitoring area is grasped; When the occurrence of a flame is detected using the initial sensitivity level, a human detection mode for detecting human voices is switched to an on setting, and a pre-trained voice recognition process is executed on the acoustic data collected by the sound collection device to determine whether or not a person is present in the monitoring area. Fire detection devices.

Images