JP3979833B2 - Automatic notification device - Google Patents

Description

[0001]
BACKGROUND OF THE INVENTION
The present invention relates to an automatic notification device that recognizes a language included in audio data and outputs a signal to the outside based on the recognized language.
[0002]
[Prior art]
Currently, automatic notification devices are used in financial institutions and convenience stores to notify the outside that an emergency such as a burglary has occurred. Many of the conventional automatic notification devices generate a signal indicating that an emergency has occurred when a store clerk operates an emergency button. The generated signal reports the occurrence of an emergency to the security company or the monitoring center via a telephone line.
[0003]
However, in recent years, the existence of the above-described conventional automatic notification device has been widely known. For this reason, it has become a concern that the criminal learns that the store clerk is going to operate the automatic notification device, so that the situation gets worse.
[0004]
For this reason, for example, a configuration has been proposed that can report the occurrence of an emergency to the outside without using an emergency button, such as the configuration described in Japanese Patent Laid-Open No. 06-131583. In the invention described in Japanese Patent Application Laid-Open No. 06-131583, a registered reporter is made to pronounce a preset keyword (for example, “Mr. Sasaki”), and this voice feature is recognized and registered as data. If the voice input by the microphone includes data that matches the registered data, the outside is notified that an emergency has occurred.
[0005]
[Problems to be solved by the invention]
However, in the case of a burglar, it is generally difficult for a person to respond calmly. For this reason, as in the invention described in Japanese Patent Laid-Open No. 06-131583, the configuration for determining the occurrence of an emergency by a specific keyword is for the user to forget or mistake the keyword, or to recognize it. There is a concern that it is not possible to report when a keyword cannot be said at a sufficient volume. In addition, there is a possibility that the robber prohibits the user from making a voice such as “Do not make noise” or “Do not speak” and cannot report.
[0006]
The present invention has been made in view of the above points, and appropriately notifies the outside that an emergency has occurred even when the user cannot correctly pronounce a keyword or cannot speak a word. An object of the present invention is to provide an automatic reporting device that can do this.
[0007]
[Means for Solving the Problems]
  In order to solve the above-described problems and achieve the object, an automatic notification device according to the first aspect of the present invention is included in voice input means for inputting voice and voice data input by the voice input means. A language recognition unit for recognizing a language, a language database in which a plurality of language groups to be compared with the language recognized by the language recognition unit are recorded, and a recognition language recognized by the language recognition unit as a result of the comparison is Language match determination means for determining whether or not the database language recorded in the language database matches, and when the language match determination means determines that the recognized language and the database language match, an internal notification Pre-notification means for outputting a signal first, and reception means for accepting pressing of the emergency push button after the signal is output by the pre-notification means; When accepting the pressing of the emergency push-button by the receiving unit, and outputs a signal to an external report destination,as well asAfter the signal is output to the internal notification destination by the pre-notification unit, if the language match determination unit determines that the recognized language and the database language match again, a signal is sent to the external notification destination. And a reporting means for outputting.
[0008]
  According to the first aspect of the present invention, a language included in speech data is recognized, and the recognized language is compared with a language database in which a plurality of language groups are recorded. As a result of the comparison, if it is determined that the recognition language matches the database language, a pre-alarm is output, and if the emergency push button is further pressed, a signal may be output to a pre-registered report destination. it can.In addition, after the pre-alarm is output, if it is determined that the language included in the audio data matches the database language, a signal can be output to a previously registered report destination.
[0025]
DETAILED DESCRIPTION OF THE INVENTION
Embodiments 1 to 3 as preferred embodiments of an automatic notification device according to the present invention will be described below in detail with reference to the accompanying drawings.
[0026]
(Embodiment 1)
FIG. 1 is a block diagram for explaining a configuration of an automatic notification device common to the first to third embodiments. The automatic notification device shown in FIG. 1 has been recognized by a speech input unit 102 for inputting speech, a language recognition unit 103 for recognizing a language included in speech data input by the speech input unit 102, and a language recognition unit 103. A language database (language DB) 108 in which a plurality of language groups contrasted with a language (recognition language) is recorded is provided.
[0027]
Further, as a result of comparison, the automatic notification device has a recognition language of language D.D. An abnormality detection unit 104 that determines whether or not it matches the language (database language) recorded in B108, and if the abnormality detection unit 104 determines that the recognition language and the database language match, an indicator lamp / buzzer message When the notification unit 106 that notifies that an abnormality has been detected using vibration or the like and the abnormality detection unit 104 determines that the recognition language and the database language match, a signal is sent to a pre-registered notification destination A reporting unit 105 for outputting and a control unit 101 for controlling the above-described configuration are provided.
[0028]
In the above configuration, the language D.I. In B108, for example, a language used to intimidate or restrict others, such as “give money”, “kill”, “do not move”, “do not make noise” is recorded. Further, for example, languages that indicate the occurrence of an emergency such as “help”, “do not kill”, and “call the police” are recorded.
[0029]
In addition, the notification unit 105 according to the first embodiment is connected to a communication line such as the telephone line 110, and can transmit an image signal captured by the security camera 109 to the report destination. It is also possible to make a phone call to the report destination using the telephone line 110 to notify the occurrence of an abnormality. Further, the voice input by the voice input unit 102 can be output to the reporting unit 105 via the control unit 101 and transmitted to the reporting destination. Further, the control unit 101 can be connected to the emergency push button 107 and can be notified by the notification unit 105 when the emergency push button 107 is operated.
[0030]
As a report destination of the report unit 105, it is conceivable to use an external organization such as a security company or the police. In addition, the notification destination of the notification unit 106 may be an individual such as a person in charge or a person in charge in a financial institution or a store, or a monitoring center in the financial institution or store that monitors an abnormality. Conceivable.
[0031]
The automatic notification device described above operates as follows. That is, the voice input unit 102 is configured using, for example, a microphone. The input voice is input to the language recognition unit 103 via the control unit 101, and the language included in the voice is recognized. Note that a specific method of language recognition performed by the language recognition unit 103 is a well-known technique, and thus further description thereof is omitted.
[0032]
The voice input unit 102 is disposed near a customer service location, for example, near a window in the case of a financial institution such as a bank or near a cashier counter in a store. Further, the voice input unit 102 is not limited to a place where a customer is served, and a plurality of voice input units 102 may be provided. Furthermore, you may attach to persons, such as a specific staff and a shop assistant.
[0033]
The abnormality detection unit 104 inputs the recognition language recognized by the language recognition unit 103, and the language D.D. Contrast with database language recorded in B108. As described above, the language D.I. In B108, a language that threatens or regulates others and a language that suggests the occurrence of an emergency (and a language that is not frequently used every day) are recorded. For this reason, the anomaly detection unit 104 contrasts the language that threatens or regulates others, the language that suggests the occurrence of an emergency and the recognition language, and the language that threatens or regulates others from the input speech, An anomaly can be detected when a language suggesting the occurrence of an emergency is recognized.
[0034]
When an abnormality is detected as described above, the control unit 101 notifies by using only a signal, a signal and sound, or a signal, sound and image, as shown in FIG. The notification using only the signal is to transmit the signal to the notification destination such as the police, the security company, or the monitoring center via the telephone line 110, for example. The report destination can identify the financial institution or store that reported from the signal source.
[0035]
Further, the notification by the signal and voice can be realized by outputting the voice input to the voice input unit 102 to the telephone line 110 as a signal together with the transmission of the signal and transmitting it to the outside. Furthermore, for notification by signal, sound, and image, the image taken by the security camera 109 is output to the telephone line 110 in the same manner as the image signal of the videophone and transmitted to the outside together with the transmission of the signal and the sound signal. Can be realized.
[0036]
According to the automatic notification device of the first embodiment, even when a user such as an employee or a store clerk cannot input a password, a language that is naturally emitted in an emergency such as “help” or “call the police” Can be detected and an abnormality can be detected. Furthermore, anomalies can be detected when the criminal utters words such as “Do not move” or “I will kill” to threaten or restrict the user. For this reason, even when the user forgets the password or is prohibited from uttering words, an abnormality can be detected and reported to the outside.
[0037]
(Embodiment 2)
Next, a second embodiment of the present invention will be described. The automatic notification device according to the second embodiment does not immediately report when the recognition language matches the database language, but can further improve the reliability of the notification by setting the notification conditions. In the following, the notification conditions set will be described as (1), (2), and (3).
[0038]
Condition (1)
According to the condition (1), the abnormality detection unit 104 determines that the recognition language matches the database language when the recognition language matches a predetermined number or more of the database languages. In this case, for example, it is not determined that the recognition language and the database language coincide with each other simply by saying that “the guard should be quiet” in order to regulate the child. On the other hand, if the criminal is “be quiet,” “make money,” or “be quick,” the recognition data will match the three database languages, and it will be determined that the recognition language matches the database language. The
[0039]
Condition (2)
According to condition (2), the abnormality detection unit 104 determines that the recognition language matches the database language when the recognition language matches a predetermined number or more of database languages included in a preset language group. To do. The language group in the condition (2) is a group of a series of languages that are supposed to be issued in an assumed emergency. As a language group, for example, “do not make noise”, “do not kill”, and “help” may be considered as one group.
[0040]
And if the condition (2) is set, for example, the guardian says “Please be quiet” to regulate the child, and if the child responds “Yes”, the recognition language and the database language match. It is not judged. On the other hand, it is determined that the recognition language and the database language match when the criminal says “Do not make noise” and the clerk responds “Do not kill” or “Help”.
[0041]
In this database language grouping, the same language may be registered in a plurality of groups. Further, a group corresponding to the place (industry) to be installed may be created and used selectively. Furthermore, group them according to the degree of risk and reliability (high possibility of an abnormal situation) and the type of abnormal situation (whether it is a burglar or just a trouble). You may make it notify with the information which shows the reliability and the kind of abnormal condition.
[0042]
Condition (3)
According to the condition (3), the abnormality detection unit 104 determines that the recognized language matches the database language when the language that matches the database language is recognized a predetermined number of times or more. To do. In this case, for example, it is not determined that the recognition language and the database language coincide with each other only when the guardian once says “keep quiet” to regulate the child. On the other hand, when the criminal repeatedly says “Please be quiet” multiple times, the recognition data matches the database language multiple times, and it is determined that the recognition language and the database language match.
[0043]
When an abnormality is detected by any of the methods described above, the reporting unit 105 reports that an abnormality has been detected using at least one of a signal, sound, and image, as in the first embodiment. In the second embodiment, the control unit 101 operates the notification unit 106 to notify, for example, a person or person in charge designated as a notification destination, a financial institution, or a monitoring center in the store.
[0044]
When the person in charge, the person in charge designated as the notification destination, or the staff of the monitoring center receives the notification, it is necessary to judge whether or not a real abnormality has occurred by visual inspection or an image taken by the security camera 109, and is necessary. In response to this, the emergency push button 107 is operated to notify at least one of the security company and the police. At this time, by designating a person who is in a position to be able to report to at least one of the security company and the police without being noticed by the criminal as a notification destination, the situation is reported as a result of calmly judging the situation. You can decide whether or not to do so.
[0045]
FIG. 3 is a flowchart for explaining processing performed by the automatic notification device that performs notification based on the notification condition described in (1). The automatic notification device first inputs a voice by the voice input unit 102 (step S301). The language recognition unit 103 recognizes a language included in the input speech (step S302), and determines whether or not the recognized recognition language matches the database (DB) language (step S303).
[0046]
If it is determined in step S303 that the recognized language matches the database language (step S303: Yes), the control unit 101 determines whether a timer (not shown) in the abnormality detection unit 104 is counting time, for example. It is determined whether or not (step S304). If the timer does not count the time as a result of the determination (step S304: No), the timer is started (step S309). Then, for example, a count n of a type counter (not shown) that counts the number of types of recognition languages that match the database language in the abnormality detection unit 104 is set to n + 1 (step S305). If it is determined in step S304 that the timer is counting (step S304: Yes), the count n of the type counter is immediately set to n + 1.
[0047]
The control unit 101 further determines whether or not the count n has reached a predetermined value N (step S306). Note that, under the condition {circle around (1)}, when the N types of recognition languages match the database language, it is determined that the recognition language matches the database language.
[0048]
If it is determined in step S306 that the count n has not reached N (step S306: No), it is determined whether or not a preset time has elapsed from the start of the count (whether or not the time is up) ( Step S310). If the time is not up as a result of the determination (step S310: No), the voice is input again to recognize the language. When the time is up (step S310: Yes), the counter is reset (step S308). Then, the process returns to step S301, and the voice is input again.
[0049]
According to such processing, when 10 minutes is set as the time to be timed up, the counter is reset when N types of recognized languages that match the database language are not recognized before the time is up. can do. Therefore, the language D. Even when N languages are recognized for some reason (to rebuke a store clerk or to restrict children), it is possible to prevent a mistaken notification from being made.
[0050]
When the count n reaches N (step S306: Yes), the reporting unit 105 outputs a signal for reporting (step S307) and reports that an abnormality has occurred. The signal to be output may be an alarm (including a message such as XX bank XX branch, an emergency has occurred) transmitted via the telephone line 110. Moreover, the signal of the audio | voice input into the audio | voice input part 102 and the image signal image | photographed by the security camera 109 may be sufficient. When the reporting unit 105 outputs a signal, the control unit 101 resets the counter (step S308). Then, the process returns to step S301, and the voice is input again.
[0051]
When a burglary occurs, the first exchange ("Do not move", "Keep quiet", "Take money", etc. and request for money) will be done in a short time, and will be done in a short time until the time is up. This time may be relatively short.
[0052]
FIG. 4 is a flowchart for explaining processing performed by the automatic notification device that performs notification based on the notification condition described in (2). When making a report based on the report conditions described in (2), the automatic notification apparatus according to the second embodiment counts that the recognition language matches the database language for each group to which the database language belongs. For example, the abnormality detection unit 104 includes a counter (not shown).
[0053]
The automatic notification device first inputs a voice by the voice input unit 102 (step S401). The language recognition unit 103 recognizes a language included in the input speech (step S402), and determines whether or not the recognized recognition language matches the database (DB) language (step S403).
[0054]
If it is determined in step S404 that the recognized language matches the database language (step S403: Yes), for example, the control unit 101 determines whether a timer (not shown) in the abnormality detection unit 104 is counting time, for example. It is determined whether or not (step S404). If the timer does not count the time as a result of the determination (step S404: No), the timer is started (step S410).
[0055]
Next, the control unit 101 selects a group to which the database language that matches the recognition language belongs (step S405). Then, the count n of the group counter dedicated to the selected group is set to n + 1 (step S406). The control unit 101 further determines whether or not the count n has reached a predetermined value N (step S407). Note that, under the condition (2), when the recognition language matches N types of database languages belonging to the same group, it is determined that the recognition language matches the database language. If it is determined in step S404 that the timer is counting (step S404: Yes), the group to which the recognized language belongs is immediately selected.
[0056]
If it is determined in step S407 that the count n has not reached N (step S407: No), it is determined whether or not a preset time has elapsed since the start of the count (whether or not time is up) ( Step S411). If the time is not up as a result of the determination (step S411: No), the voice is input again to recognize the language.
[0057]
When the time is up (step S411: Yes), the counter is reset (step S409). And it returns to step S401 and inputs a sound again. According to such processing, when, for example, 10 minutes is set as the time to be timed up, N recognition languages that match the database language belonging to the same group have not been recognized by the time up. You can reset the counter. Therefore, even if N languages are recognized for a certain reason (to rebuke a store clerk or to restrict children) for a long time, it is possible to prevent an erroneous report from being made. be able to.
[0058]
When the count n reaches N (step S407: Yes), the reporting unit 105 outputs a signal for reporting (step S408) and reports that an abnormality has occurred. The signal to be output may be an alarm (including a message such as XX bank XX branch, an emergency has occurred) transmitted via the telephone line 110. Moreover, the signal of the audio | voice input into the audio | voice input part 102 and the image signal image | photographed with the security camera may be sufficient. When the reporting unit 105 outputs a signal, the control unit 101 resets the counter (step S409). Then, the process returns to step S401, and the voice is input again.
[0059]
FIG. 5 is a flowchart for explaining processing performed by the automatic notification device that performs notification based on the notification condition described in (3). When making a report based on the notification conditions described in (2), the automatic notification apparatus of the second embodiment counts the number of times that the recognized recognition language matches the database language (not shown). For example, in the abnormality detection unit 104.
[0060]
The automatic notification device first inputs a voice by the voice input unit 102 (step S501). The language recognition unit 103 recognizes a language included in the input speech (step S502), and determines whether or not the recognized recognition language matches the database (DB) language (step S503).
[0061]
If it is determined in step S503 that the recognition language matches the database language (step S503: Yes), the control unit 101 determines whether a timer (not shown) in the abnormality detection unit 104 is counting time, for example. It is determined whether or not (step S504). If the timer does not count the time as a result of the determination (step S504: No), the timer is started (step S509). Then, the count n of the above-mentioned number counter is set to n + 1 (step S505). The control unit 101 further determines whether or not the count n has reached a predetermined value N (step S506). Under the condition (3), when one recognition language matches the database language N times, it is determined that the recognition language matches the database language. If it is determined in step S504 that the timer is counting (step S504: Yes), the count n of the type counter is immediately set to n + 1.
[0062]
In step S506, when it is determined that the count n has not reached N (step S506: No), it is determined whether or not a preset time has elapsed from the start of the count (whether or not the time is up) ( Step S510). If the time is not up as a result of the determination (step S510: No), the voice is input again to recognize the language.
[0063]
When the time is up (step S510: Yes), the counter is reset (step S508). Then, the process returns to step S501 and the voice is input again. According to such processing, when, for example, 10 minutes is set as the time up, the counter is reset when one recognition language does not match the database language N times before the time up. be able to. Therefore, even if one recognized language that matches the database language is recognized N times for some reason (to rebuke a store clerk or to restrict a child), it is prevented that an erroneous report is made. Can do.
[0064]
When the count n reaches N (step S506: Yes), the reporting unit 105 outputs a signal for reporting (step S507) and reports that an abnormality has occurred. The signal to be output may be an alarm (including a message such as XX bank XX branch, an emergency has occurred) transmitted via the telephone line 110. Moreover, the signal of the audio | voice input into the audio | voice input part 102 and the image signal image | photographed with the security camera may be sufficient. When the reporting unit 105 outputs a signal, the control unit 101 resets the counter (step S508). Then, the process returns to step S501 and the voice is input again.
[0065]
In the second embodiment described above, the recognition language matches a plurality of types of database languages, matches a plurality of types of database languages belonging to the same group, and further, one recognition language matches a database language multiple times. It is determined that the recognition language matches the database language. For this reason, the same effect as obtained in Embodiment 1 can be obtained, and the reliability of abnormality detection can be improved.
[0066]
(Embodiment 3)
FIG. 6 is a diagram for explaining notification processing according to the third embodiment. When an abnormality is detected by any of the conditions (1), (2), and (3) shown in the second embodiment, the control unit 101 issues a pre-alarm.
[0067]
For example, in the case of a bank, a pre-alarm is issued to the person at a position away from the branch manager or the customer service counter by the notification unit 106 to notify that an abnormal situation may occur at the customer service counter. Branch managers etc. calmly judge the situation from a remote location, and if it is judged abnormal, operate the emergency push button, police or guard by using only the signal, signal and voice, signal and voice and image Report to a company or other company. In this case, since the other person who receives the pre-alarm is located away from the criminal, it is difficult for the criminal to see even if the emergency push button is operated.
[0068]
In addition, if an abnormality is detected again under any of the conditions (1), (2), or (3) with the pre-alarm issued, it is considered that the possibility of an abnormal situation is sufficiently high, Report only by signal, signal and voice, signal and voice and image. By doing in this way, even if the person who receives the pre-alarm is away from the seat and the person who receives the pre-alarm is in a situation where the emergency push button cannot be operated, the notification can be made.
[0069]
In the third embodiment described above, when an abnormality is detected, the third party is first notified to the inside that there is a possibility that an abnormal situation has occurred. Whether or not to make a report can be determined based on a calm judgment by the person, and a highly reliable report can be made.
[0070]
In addition, banks and other financial institutions have prepared manuals for dealing with incidents, and training on how to deal with incidents when incidents occur. If there is a risk, they may not be fully utilized. However, by issuing a pre-alarm, it can be expected to provide time for the store clerk to be prepared and to always take a calm action.
[0071]
Further, even when the notification is not made by another person inside, if the abnormality is detected again, it is possible to notify that the possibility of the occurrence of the abnormal situation is sufficiently high.
[0072]
In addition, it is possible to notify only the inside person shown in the third embodiment and use only the part of the notification by operating the emergency push button by the other person who has received the notification.
[0073]
In addition, this invention is not limited to an above-described structure, For example, it can also comprise as follows. That is, the conditions set in the second embodiment are not limited to the above conditions (1), (2), and (3). For example, the database language is given a weight, and the recognition language is set. Count the database language weights that match. Then, when the total value of the count reaches a predetermined value, an abnormality may be detected and an alarm may be issued.
[0074]
Further, the present invention is not used only for the intrusion of a burglar, but can be applied to an apparatus for inputting a word uttered by an employee and checking the customer service attitude, for example. FIG. 7 is a block diagram for explaining the configuration of the automatic notification device of the present invention when used for checking the customer service attitude. In the automatic notification device shown in FIG. 7, the recording unit 701 that records that the control unit 101 determines that the recognition language matches the database language is used as a notification destination.
[0075]
For example, the language D.I. In B702, a basic greeting such as “Thank you for coming” or “Thank you” or a language that is instructed to be repeated in a campaign or the like is memorized. Also, for example, the language D.I. In B702, a rough word for a customer such as “Do you want to eat” or a wrong word such as “Our company” is stored as a database language.
[0076]
The language recognition unit 103 recognizes the language included in the voice input unit 102, and the match detection unit 703 determines whether the recognized language matches the database language. If it is determined that they match, a signal indicating that the recognized language matches the database language is sent to the recording unit 701. The recording unit 701 records the contents of the report (such as the matched database language, the report time, the number of times).
[0077]
The manager of customer service contrasts the data recorded in the recording unit 701 with the records of the person in charge who responds to the customer, and who is in charge of greetings such as “Thank you” or “Thank you”. You can know if you are or are neglected. Or you can know which person in charge is using violent language or using the wrong language. That is, the record recorded in the recording unit 701 can be a material when the manager instructs the person in charge about the customer service attitude.
[0078]
By the way, the situation where a burglar or the like invades generally does not occur frequently. As described above, the use of the automatic notification device of the present invention for checking the language has an effect that it is possible to check whether the automatic notification device installed over a long period of time functions correctly.
[0079]
【The invention's effect】
  As described above, the invention according to claim 1 recognizes the language included in the audio data, and when it is determined that the recognized language matches the database language, the pre-alarm is output. Furthermore, when the emergency push button is pressed, a signal can be output to a previously registered report destination. For this reason, the pre-alarm is issued to the person at a position away from the branch manager or the customer service counter by the notification unit to notify that there is a possibility that an abnormal situation has occurred at the customer service counter. The branch manager etc. calmly judges the situation from a remote location, and when it is judged abnormal, operates the emergency push button and reports to a reporting destination such as the police or a security company by a signal. In this case, since the other person who receives the pre-alarm is at a position away from the criminal, there is an effect that even if the emergency push button is operated, the criminal is hardly noticed.In addition, if it is determined that the language included in the audio data matches the database language after the pre-alarm is output, a person who receives the pre-alarm by outputting a signal to a pre-registered report destination Even if the person who is away from the seat or the person who receives the pre-alarm is in a situation where the emergency push button cannot be operated, it is possible to report.
[Brief description of the drawings]
FIG. 1 is a block diagram for explaining a configuration of an automatic notification device common to Embodiments 1 to 3 of the present invention.
FIG. 2 is a diagram for explaining a notification form according to the first embodiment;
FIG. 3 is a flowchart for explaining processing performed in the second embodiment;
FIG. 4 is another flowchart for explaining processing performed in the second embodiment.
FIG. 5 is another flowchart for explaining processing performed in the second embodiment;
FIG. 6 is a diagram for explaining a notification form according to the third embodiment.
FIG. 7 is a diagram for explaining another configuration of the automatic notification device of the present invention.
[Explanation of symbols]
101 Control unit
102 Voice input unit
103 Language recognition unit
104 Anomaly detector
105 Report Department
106 Notification unit
107 Emergency push button
108,702 Language Database
109 Security Camera
110 Telephone line
701 Recording unit
703 Match detection unit

Claims (1)

Voice input means for inputting voice;
Language recognition means for recognizing a language included in the voice data input by the voice input means;
A language database in which a plurality of language groups contrasted with the language recognized by the language recognition means are recorded;
As a result of the comparison, a language match determination means for determining whether or not the recognition language recognized by the language recognition means matches the database language recorded in the language database;
When it is determined by the language match determination means that the recognition language and the database language match, pre-notification means for outputting a signal to an internal report destination;
An acceptance means for accepting pressing of an emergency push button after a signal is output by the pre-notification means;
When the reception unit accepts the pressing of the emergency push button, a signal is output to an external notification destination, and the signal is determined again after the pre-notification unit outputs a signal to the internal notification destination. If it is determined by the means that the recognized language and the database language match, the reporting means for outputting a signal to an external reporting destination;
An automatic notification device characterized by comprising:

Images