WO2019026117A1 - Security system - Google Patents

Abstract

The camera (12) of the security system (10) comprises an imaging unit (12a) for imaging the surroundings of the camera (12), and a person detection unit (for detecting a person based on an image imaged by the imaging unit (12a) 12c), a characteristic portion detection unit (12d) for detecting a characteristic portion of the person from the image of the person detected by the person detection unit (12c), and an image of the characteristic portion detected by the feature portion detection unit (12d) And a person identification unit (12e) for identifying a person based on the

Description

Security system

The present invention relates to a security system having at least a camera for imaging a person.

For example, a security system that transmits an image captured by a camera installed in a building or the like to a server of an administrator is disclosed in JP-A-2009-98814, JP-A-2013-30116, and JP-A-2015-2506. It is disclosed in the official gazette.

Also, in recent security systems, images captured by a camera are distributed to portable devices such as smartphones via a communication line, and further reflected in the images based on changes in luminance of the images captured by the camera. It is also performed to detect a person and to start recording of a captured image triggered by the detection result.

However, in the conventional security system, for example, when an image of a visitor visiting a building such as a house is imaged by a camera, the brightness of the movement of a non-human moving body reflected in the imaged image or the brightness of the illumination device In some cases, all changes can be regarded as human movements. Thus, it can not be appropriately determined whether the moving object or the like reflected in the image is a visitor or what kind of person the visitor is. As a result, when the resident of the house is absent, it is not possible to provide an appropriate service (for example, interaction with the visitor, delivery of the image of the visitor) tailored to the visitor.

The present invention has been made in consideration of such problems, and an object of the present invention is to provide a security system capable of providing an appropriate service tailored to a person captured by a camera.

The present invention relates to a security system having at least a camera for imaging a person.

Further, in order to achieve the above object, in the present invention, the camera includes an imaging unit for imaging the periphery of the camera, a person detection unit for detecting the person based on an image imaged by the imaging unit, A feature portion detection unit for detecting a feature portion of the person from the image of the person detected by the person detection unit; a person identification unit for identifying the person based on the image of the feature portion detected by the feature portion detection unit; Have.

As described above, by mounting an artificial intelligence on the camera and causing the person detection unit, the feature portion detection unit, and the person identification unit to function, provision of services tailored to a specific person such as a visitor can be performed locally. It is possible to do on the camera side as a region. Thereby, for example, even when the resident of the house where the camera is installed is absent, the camera can take an appropriate response to the visitor. As a result, the convenience of the security system can be improved.

In this case, the person detection unit detects a moving object reflected in the image, or detects a change in noise of the image, or a change in brightness of the image in the image captured by the camera. The person may be detected by detecting

Thus, the person can be easily detected using only the image captured by the imaging unit. As a result, with the detection result of the person detection unit as a trigger, the characteristic part detection unit and the person identification unit can execute each process.

In addition, the camera may further include a first microphone that detects sound around the camera. In this case, the person detection unit may detect the person based on the image captured by the imaging unit and the sound detected by the first microphone.

As described above, by using the image captured by the imaging unit and the sound detected by the first microphone in combination, it is possible to detect the person with high accuracy.

Furthermore, when the camera images the periphery of the camera continuously or at predetermined time intervals, the characteristic part detection unit detects the characteristic part from the image of the person each time the person detection unit detects the person The feature portion may be tracked by detecting the

This makes it possible to easily grasp the behavior of the person and to take an appropriate response to the person.

Further, the camera may further include a person information registration unit in which person information on a predetermined person is registered. In this case, the person specifying unit extracts a feature amount of the feature portion from the image of the feature portion detected by the feature portion detecting unit, and a feature portion of the person extracted by the feature amount extracting unit A similarity degree determination processing unit that determines the similarity between the feature amount of the person and the feature amount of the feature portion of the person indicated by the person information; the feature amount of the feature portion of the person extracted by the feature amount extraction unit; And an age / sex estimation processing unit that estimates the age and / or sex of the person detected by the person detection unit.

This makes it possible to easily and efficiently specify what kind of person the person detected by the person detection unit is.

In this case, the characteristic portion may be at least a face or appearance of a person detected by the person detection unit. As a result, when the camera is installed in a building such as a house, it is possible to easily detect a visitor to the building or a movement of a resident entering or leaving the building. In addition, the appearance includes, for example, the appearance of the front side or the side of the visitor, and the back appearance of a resident coming out of the building.

In addition, the feature portion may include clothes of the person, patterns, colors or patterns of the clothes, or marks attached to the clothes. Thereby, when the camera is installed in a building such as a house, it is possible to easily identify what kind of person is a visitor to the building or a person who goes in and out of the building.

The camera may further include a first response determination unit that determines a response to the person according to the identification result when the person identification unit identifies the person.

Thereby, it is possible to take an appropriate response according to the person image of the person.

Specifically, in accordance with the determination result of the first response determination unit, the security system executes the following first to third measures for the person.

The first action is to execute a dialogue between the person and the camera, and the artificial intelligence function mounted on the camera corresponds to the person only by the internal processing of the camera. .

That is, when the camera determines that the scenario registration unit in which a plurality of scenarios are registered and the first response determination unit execute the dialogue with the person based on the specification result, the camera registration unit A first dialog engine unit that selects a scenario according to a dialog; a first speaker that outputs the scenario selected by the first dialog engine unit as a voice to the outside; and a second microphone that detects the voice of the person Furthermore, it has.

This allows the camera to easily interact with the person according to the scenario. In addition, since the first dialogue engine unit, which is a function of artificial intelligence, is provided, even if the amount of image and voice data increases due to the execution of the dialogue, the first dialogue engine unit can be easily and It becomes possible to process promptly. Furthermore, since the first dialogue engine unit performs dialogue learning through dialogue with the person, it is possible to improve various functions of the artificial intelligence including the dialogue function.

In this case, the camera may further include a related information registration unit in which information on the installation place of the camera, information on the periphery of the installation place, and / or information on a predetermined person is registered as related information. . Then, when the first response judging unit determines execution of a dialog with the person specified by the person specifying unit based on the specifying result, the first dialogue engine unit determines the relation from the related information registration unit Information may be extracted, and a scenario corresponding to the extracted related information and the dialogue may be selected from the scenario registration unit.

Thereby, referring to the related information, it is possible to select a more appropriate scenario and improve the accuracy of the interaction with the person. In addition, when the camera is installed in a building such as a house, if a visitor to the building is known in advance, the related information of the visitor may be registered in the related information registration unit. When a person visits the building, an appropriate dialogue can be performed for the visitor.

In this case, the camera may further include a related information collection unit that collects the related information from the outside and registers the collected related information in the related information registration unit. Thus, since the latest related information is registered in the related information registration unit, it is possible to further improve the accuracy of the dialog with the person.

In the first dialogue engine unit, the imaging unit captures an image of the person when the first response determination unit determines execution of a dialog with the identified person based on the identification result. The relevant information collected by the relevant information collection unit may be extracted from the relevant information registration unit at the time point of occurrence, and the scenario corresponding to the extracted relevant information and the dialogue may be selected from the scenario registration unit. This further improves the accuracy of the dialogue with the person.

Furthermore, when the related information according to the person specified by the person specifying unit is not registered in the related information registration unit, the first dialogue engine unit is another related information registered in the related information registration unit. Information may be extracted, and the extracted other related information and a scenario corresponding to the interaction may be selected from the scenario registration unit and / or notified to the outside. Thereby, when the camera is installed in a building such as a house, even when the visitor to the building and the visitor indicated by the related information registered in the related information registration unit do not match, Other relevant information can be used to perform the interaction with the visitor. In addition, it is possible to notify a resident who is away from the building by reporting the visit of the visitor.

The second response is to notify the holder of an external portable device (for example, a resident of a house where the camera is installed) from the camera by notifying the presence of the person (for example, a visitor's visit). is there.

That is, the security system further includes a portable device connectable to the camera via a communication line. In this case, the camera further includes a first communication unit capable of communicating with at least the portable device via the communication line. Then, when the first response determination unit determines to notify the portable device of the presence of the person based on the identification result, the first communication unit transmits the notification to the portable device via the communication line. , Notify the existence of the person.

As a result, the holder of the portable device in the absence of the installation location of the camera can receive the notification in real time even if not at the installation location. Thereby, the security system can provide appropriate services to the holder.

Also, in the second response, after (the holder of) the portable device receives the notification, it is possible to interact with the person.

Specifically, the portable device can communicate with at least the camera via the communication line, a second communication unit, a display unit, an operation unit, a second speaker, and a third microphone. Have.

Here, when the second communication unit receives the notification, the display unit displays the notification and / or the second speaker outputs the notification as a sound. On the other hand, when the holder of the portable device operates the operation unit and instructs start of a dialogue with the person, the second communication unit transmits the first communication unit via the communication line. Sends an instruction signal instructing the start of the dialogue.

In this case, when the first communication unit receives the instruction signal, an image of the person captured by the imaging unit and voice data of the person detected by the second microphone are transmitted through the communication line. It transmits to said 2nd communication part via.

Thereby, the display unit displays the image of the person received by the second communication unit, and the second speaker outputs the voice data of the person received by the second communication unit as sound, and The second communication unit transmits voice data of the holder detected by the third microphone to the first communication unit via the communication line. As a result, the first speaker outputs the sound data of the holder received by the first communication unit as a sound.

In this way, since the dialogue is started between the person on the camera side and the holder of the portable device, the holder is suitable in real time even if he is not at the installation location of the camera. Service (interactive service). Thereby, the convenience of the security system can be further improved. Even in this case, the first dialogue engine unit can perform dialogue learning through the dialogue with the person and improve various functions of the artificial intelligence including the dialogue function.

Then, in the first response or the second response, when the dialog with the person ends, the camera performs the following processing.

That is, after the end of the dialog with the person, the camera records at least the image of the person captured by the imaging unit during the dialog and the voice data of the person detected by the second microphone. It further has a data storage unit. In this case, after the end of the dialogue, the first dialogue engine unit creates a new scenario based on the contents of the dialogue with the person, and registers the created new scenario in the scenario registration unit.

As a result, it is possible to confirm at a later date what kind of dialogue has been performed with the person, and the convenience of the security system can be further improved. Further, since the artificial intelligence is mounted on the camera, the first dialogue engine unit processes the data appropriately and promptly even if the amount of data stored in the data storage unit increases dramatically. The new scenario can be created and registered in the scenario registration unit. Moreover, since the first dialogue engine unit creates the new scenario by dialogue learning through dialogue with the person, it is easy to select an optimal scenario each time the dialogue with the person is executed. That is, since the new scenario is created and registered in the scenario registration unit each time the user interacts with the person, the accuracy of the artificial intelligence function mounted on the camera is improved, and the optimal scenario is selected. Dialogue can be performed.

In addition, in the third correspondence, when it is necessary to specify the person in detail, for example, when it is difficult to specify the person with the artificial intelligence function on the camera side, the external server may It is a response to inquire (request) the detailed identification of the person.

Specifically, the security system is connected to a server connectable to the camera via a communication line, and connected to the server, and includes information on an installation location of the camera and peripheral surroundings including information on the periphery of the installation location. It further has a peripheral information database in which information is registered, and a personal information database connected to the server and in which personal information of a predetermined person is registered.

In this case, when the first response determination unit determines to inquire the server of the detailed identification of the person detected by the person detection unit on the basis of the identification result, the first response determination unit determines the detail of the person via the communication line. Query the server for identification.

On the other hand, the server includes: a third communication unit that receives the content of the inquiry from the first response determination unit; an inquiry content determination unit that determines the content of the inquiry received by the third communication unit; The peripheral information is acquired from the peripheral information database via the third communication unit based on the determination result of the above, and the personal information corresponding to the person indicated by the content of the inquiry is acquired from the personal information database and acquired And a data matching unit for specifying the person in detail by matching the peripheral information and the personal information with the content of the inquiry.

Thus, the peripheral information database and the personal information database are provided on the server side, and the amount of data required to identify the person is much larger than that of the camera. Therefore, by matching the peripheral information and the personal information with the contents of the inquiry, the server can specify the person's image etc. in detail. Further, by installing artificial intelligence in the server and making it function as the inquiry content judgment unit and the data matching unit, it is possible for the server as a local area to provide services tailored to the person. Become. Even in this case, even if there is no resident of the house where the camera is installed, the server can take an appropriate response to the person. As a result, the convenience of the security system can be improved. Furthermore, by matching the data, it is possible to easily associate the person with the content such as the surrounding information and the personal information.

In addition, the security system may further include a scenario database connected to the server and registered with a plurality of scenarios. In this case, when the data matching unit specifies a person indicated by the contents of the inquiry in detail, the server determines a response to the person according to the specification result, and the second response determining unit; and the second response A second dialogue engine unit that selects a scenario corresponding to the dialogue from the scenario database via the third communication unit when the judgment unit decides to execute dialogue with the person based on the identification result; In addition, The camera further includes a second speaker for outputting the scenario selected by the second dialogue engine unit as a voice to the outside, and a third microphone for detecting the voice of the person.

Thus, the camera can easily and accurately interact with the person according to the scenario selected on the server side. In addition, since the server has the second dialogue engine unit which is one function of artificial intelligence, even if the amount of image and voice data increases due to the execution of the dialogue, the second dialogue engine unit It becomes possible to process easily and promptly. Furthermore, the second dialogue engine unit can perform dialogue learning through dialogue with the person to improve various functions of the artificial intelligence including the dialogue function.

Further, the security system is connected to the server, and after the end of the dialog with the person, at least the image of the person captured by the imaging unit during the dialog and the person detected by the third microphone It further has a dialogue history database in which speech data is recorded as dialogue history. In this case, after the end of the dialog, the second dialog engine unit creates a new scenario using the dialog history recorded in the dialog history database, and registers the created new scenario in the scenario database. Do.

As a result, it is possible to confirm at a later date what kind of dialogue has been performed with the person, and the convenience of the security system can be further improved. Further, since the artificial intelligence is mounted on the server, the second dialogue engine unit processes the data appropriately and promptly even if the amount of data accumulated in the dialogue history database increases dramatically. , Can create the new scenario. Moreover, since the new scenario is created using the dialogue history accumulated in the dialogue history database, the second dialogue engine unit facilitates an optimal and latest scenario when interacting with the person. It is possible to select That is, since the second dialogue engine unit creates the new scenario by dialogue learning through dialogue with the person, it is easy to select an optimal scenario each time the dialogue with the person is executed. That is, even in the third response, the new scenario is created each time the user interacts with the person, and is registered in the scenario database, so that the accuracy of the artificial intelligence function installed in the server is improved, which is optimal. Dialog can be selected by selecting

It is a block diagram of a security system concerning this embodiment. Figure 2 is a detailed block diagram of the camera of Figure 1; Figure 2 is a detailed block diagram of the mobile device and server of Figure 1; It is a flowchart which shows operation | movement of the security system of FIG. It is a flowchart which shows operation | movement of the security system of FIG. It is a flowchart which shows operation | movement of the security system of FIG. It is a flowchart which shows operation | movement of the security system of FIG.

The present invention will be described in detail below with reference to preferred embodiments and with reference to the attached drawings.

[1. Configuration of this embodiment]
The configuration of the security system 10 according to the present embodiment will be described with reference to FIGS. 1 to 3.

The security system 10 monitors a visitor (person) who visits a monitoring target such as a building including a house, a land, an office, etc., a person who enters and leaves the monitoring target (for example, a resident who enters and leaves the housing) It applies to services such as security.

The security system 10 is installed, for example, at an entrance of a monitoring target, and a camera 12 for capturing an image of the periphery of an installation location including a visitor and a person who enters or leaves the monitoring target 16 and a portable device 18 such as a smartphone which is owned by the owner of the monitoring target (for example, a resident of a house) and connected to the camera 12 and the server 16 via the communication line 14.

Here, as an example, when the monitoring target is a house and a visitor (person) visits the entrance of the house from the outside or a resident (person) other than the owner of the portable device 18 enters and leaves the house explain. Therefore, the person to be imaged by the camera 12 is not limited to an external person who visits the monitoring object, and the person concerned with the monitoring object who does not possess the portable device 18 (for example, possesses the portable device 18 It is a concept that also includes the family of residents. Furthermore, in FIG. 1, the case where the entrance to a house is monitored in a state in which signals can be transmitted and received by wireless communication with the camera 12, the server 16, and the portable device 18 via the communication line 14 will be described.

The server 16 is a cloud server of a service provider such as monitoring or security, and in the cloud 20 in which the server 16 is included, the installation location of the camera 12 and the periphery where information around the installation location is stored. A personal information database 24 (hereinafter also referred to as a personal information DB 24) in which personal information of a plurality of persons including a visitor is stored, an information database 22 (hereinafter also referred to as a surrounding information DB 22), a visitor and a camera 12 Scenario database 26 (hereinafter, also referred to as scenario DB 26) in which the scenario used in executing the dialogue with the user is stored, and the history database 28 as a dialogue history database in which the dialogue history between the visitor and the camera 12 is recorded. (Hereinafter, also referred to as history DB 28) are arranged.

The peripheral information is the installation location of the camera 12 and weather information, traffic information, crime prevention information and the like around the installation location, and is stored in the peripheral information DB 22 in association with the time when these pieces of information are acquired. Personal information is information that identifies a person such as a visitor or a resident of a house, such as a person's photo, name, address, work place, work place address, and the like. In addition, personal information may include personal information which is information on a predetermined person described later. It should be noted that the server 16 and each of the DBs 22 to 28 have functions that complement each other in the cloud 20.

As shown in FIG. 2, the camera 12 includes an imaging unit 12a, a microphone 12b (first and second microphones), a person detection unit 12c, a feature detection unit 12d, a person identification unit 12e, and a response determination unit 12f (first Response determination unit), dialogue engine unit 12g (first dialogue engine unit), related information collection unit 12h, communication unit 12i (first communication unit), speaker 12j (first speaker), memory 12k, database 12l (person information registration) , A scenario registration unit, a related information registration unit), and a data storage unit 12m. The person specifying unit 12e includes a feature amount extracting unit 12n, a similarity degree determination processing unit 12o, and an age / sex estimation processing unit 12p.

The camera 12 reads out and executes the program stored in the memory 12k, whereby the person detection unit 12c, the feature portion detection unit 12d, the person specification unit 12e (feature amount extraction unit 12n, similarity determination processing unit 12o, age, The functions of the gender estimation processing unit 12p), the response determination unit 12f, the dialogue engine unit 12g, and the related information collection unit 12h are realized. The camera 12 is a camera equipped with an artificial intelligence function, and is implemented by execution of a program. The person detection unit 12c, the characteristic part detection unit 12d, the person identification unit 12e, the response determination unit 12f, and the dialogue engine unit 12g The related information collection unit 12 h is responsible for the artificial intelligence function.

The imaging unit 12 a captures an area around the installation location of the camera 12. The microphone 12 b detects sound around the installation location of the camera 12 (for example, voice of a person captured by the camera 12). The human detection unit 12c detects a person who is captured in an image based on an image captured by the imaging unit 12a and further audio data detected by the microphone 12b. The feature part detection unit 12d detects a feature part of the person from the image of the person detected by the person detection unit 12c.

The characteristic part of the person is at least the face or appearance of the person detected by the person detection unit 12c. The appearance includes the appearance when the person is viewed from the front, the side or the side, and the back of the person. In addition, the feature portion may include the clothes of the person detected by the person detection unit 12 c, the pattern, color or pattern of the clothes, or a mark attached to the clothes of the person. For example, if the person visits the installation location of the camera 12 (for example, various delivery companies, home delivery companies, postal companies (the delivery person in charge or sales person in charge)), the person wears them. The characteristic color or pattern of the uniform or hat to be used, the sign (trademark) of the business operator attached to the uniform or hat, or the like is the characteristic part of the person. These characteristic portions are an example, and for a person reflected in an image captured by the camera 12, a portion capable of specifying the person may be a characteristic portion of the person.

The person specifying unit 12e specifies the person detected by the person detecting unit 12c based on the image of the characteristic portion detected by the characteristic portion detecting unit 12d. Specifically, the feature quantity extraction unit 12 n extracts the feature quantity of the feature part from the image of the feature part detected by the feature part detection unit 12 d. The similarity determination processing unit 12 o is similar to the extracted feature amount and the feature amount of the feature portion of the person included in the person information on the predetermined person registered in the database 12 l (hereinafter, also referred to as DB 12 l). Determine the degree. The age / sex estimation processing unit 12 p estimates the age and / or gender of the person by comparing the feature amount of the feature portion with the person information. The personal information is information for specifying a predetermined person, and includes information including the name, the age, the gender, the image of the characteristic portion, the feature amount, and the like of the predetermined person.

When the person identification unit 12 e identifies a person, the response determination unit 12 f determines the response to the person according to the identification result.

When the response judging unit 12f decides to execute the dialog with the person based on the identification result of the person specifying unit 12e, the dialog engine unit 12g conforms to the dialog among the plural scenarios registered in the DB 12l. Select a scenario to The scenario means a predetermined fixed sentence indicating the contents of the conversation thrown from the camera 12 to the person when the camera 12 and the person interact with each other. In addition, in the DB 12l, it is desirable that a plurality of scenarios be classified into predetermined categories and registered in accordance with the type of person, dialogue, and the like.

The related information collection unit 12 h uses, as related information, information on the installation location of the camera 12, information on the periphery of the installation location, and / or information on a predetermined person (person information described above) as the communication information 14 and the communication unit 12 i Collected from the outside, and registered the collected related information in DB 12 l. For example, the related information collection unit 12 h collects peripheral information stored in the peripheral information DB 22 and personal information stored in the personal information DB 24 as related information via the communication line 14 and the communication unit 12 i. It is preferable that the related information collection unit 12 h register the related information in the DB 12 l in association with the time when the related information was collected.

The communication unit 12i transmits and receives signals by wireless communication with the portable device 18 and / or the server 16 via the communication line 14. The speaker 12 j outputs as a voice the fixed phrase of the scenario selected by the dialogue engine unit 12 g. After the end of the dialog with the person, at least the image of the person captured by the imaging unit 12a and the voice data of the person detected by the microphone 12b are recorded in the data storage unit 12m. In addition, after the end of the dialogue, the dialogue engine unit 12g creates a new scenario according to the contents of the dialogue, and registers the created new scenario in the DB 12l. In this case, new scenarios are distributed and registered in predetermined categories according to the type of person, dialogue, and the like.

As shown in FIG. 3, the server 16 includes a communication unit 16 a (third communication unit), an inquiry content judgment unit 16 b, a data matching unit 16 c, a response judgment unit 16 d (second response judgment unit), and a dialogue engine unit 16 e (first 2) A dialogue engine unit) and a memory 16f. On the other hand, the portable device 18 includes a communication unit 18a (second communication unit), a control unit 18b, a memory 18c, a display unit 18d, an operation unit 18e, a speaker 18f (second speaker), and a microphone 18g (third microphone).

The server 16 reads out and executes the program stored in the memory 16f to realize the functions of the inquiry content determination unit 16b, the data matching unit 16c, the response determination unit 16d, and the dialogue engine unit 16e. The server 16 is a computer equipped with an artificial intelligence function, and the inquiry content judgment unit 16b, the data matching unit 16c, the response judgment unit 16d, and the dialogue engine unit 16e realized by execution of a program have artificial intelligence functions. I am responsible.

The communication unit 16 a of the server 16 transmits and receives signals by wireless communication with the camera 12 and / or the portable device 18 via the communication line 14. When the communication unit 16a receives the inquiry content from the response determination unit 12f of the camera 12, the inquiry content determination unit 16b determines what kind of inquiry content has been received. The inquiry content is an inquiry for requesting identification of the person in more detail on the server 16 side although the person identification has been performed on the camera 12 side. For example, even if it is difficult to specify a person taken by the imaging unit 12a with the person specifying unit 12e with high accuracy, the above-described inquiry content is transmitted from the camera 12 to the server 16.

The data matching unit 16c acquires the peripheral information of the camera 12 from the peripheral information DB 22 through the communication unit 16a based on the determination result of the inquiry content determination unit 16b, and personal information according to the person from the personal information DB 24. By matching the acquired peripheral information and personal information with the contents of the inquiry, the person is specified in detail.

When the data matching unit 16c specifies a person in detail, the response determination unit 16d determines the correspondence to the person according to the specification result. When the response determination unit 16 d decides to execute the dialog with the person based on the identification result in the data matching unit 16 c, the dialog engine unit 16 e responds to the scenario from the scenario DB 26 via the communication unit 16 a. Select The scenario stored in the scenario DB 26 is the same as the scenario stored in the DB 12 l of the camera 12. However, it should be noted that the amount of scenario data accumulated in the scenario DB 26 is much larger than the amount of scenario data accumulated in the DB 12 l. Further, in the scenario DB 26 as well, it is desirable that each scenario be classified and registered in a predetermined category according to the type of person, dialogue, and the like.

On the other hand, the communication unit 18 a of the portable device 18 transmits and receives signals by wireless communication with the camera 12 and / or the server 16 via the communication line 14. The control unit 18 b controls each unit in the portable device 18 by reading and executing the program stored in the memory 18 c. The display unit 18 d is, for example, a display portion of a touch panel, and displays desired information. The operation unit 18 e is, for example, a soft key of a touch panel, and a holder of the portable device 18 can issue a desired instruction to the portable device 18 by operating the operation unit 18 e. The speaker 18 f outputs audio data as sound to the outside. The microphone 18 g detects the voice of the holder of the portable device 18.

[2. Operation of this embodiment]
Next, the operation of the security system 10 according to the present embodiment will be described with reference to FIGS. 4 to 7. The description of the operation will be made with reference to FIGS. 1 to 3 as necessary. Here, the following three embodiments (first to third embodiments) and modified examples of the first to third embodiments will be described in order.

<2.1 First Embodiment>
A first embodiment (first correspondence) will be described with reference to FIG. In the first embodiment, the person specifying unit 12e specifies a person from the image captured by the imaging unit 12a, and based on the specification result, the camera 12 and the person interact with each other. Therefore, in the first embodiment, all the processing is performed in the camera 12, and information between the camera 12 and the portable device 18 and / or the server 16 via the communication line 14 can be Note that no transmission or reception takes place. In the first embodiment of FIG. 4, a case where a visitor as a person visits a house will be described as an example.

In step S1 of FIG. 4, when the visitor comes to the entrance of a house in which the camera 12 is installed, the imaging unit 12a of the camera 12 images around the installation location of the camera 12 including the visitor. In this case, the camera 12 may image the periphery of the installation location before the visitor visits, or remains in a standby state until the visitor visits, and starts imaging when the visitor visits May be Also, the microphone 12 b detects a sound around the installation location of the camera 12 including the voice of the visiting visitor.

The human detection unit 12c detects a visitor based on the image captured by the imaging unit 12a. Specifically, the person detection unit 12c detects a visitor by the following method.

When the imaging unit 12a images the periphery of the camera 12 continuously or at predetermined time intervals, that is, when the imaging unit 12a is performing moving image shooting or intermittent still image shooting, the human detection unit 12c is a camera 12 Detects, for images taken continuously or at predetermined time intervals, a moving object reflected in the image, a change in noise in the image, or a change in brightness of the image By detecting visitors.

That is, when the resident of the house (the owner of the portable device 18) is out, and the house is absent, the moving object is in spite of the fact that some movement should not occur even if the image is taken by the camera 12. When it is detected, it can be determined that the visitor has visited.

In addition, when another resident (family) of the house returns home, the brightness of the image captured by the camera 12 changes if the other resident lights the light in the house. Even in this case, it can be determined that another resident who is a visitor has visited.

Furthermore, in addition to the image captured by the camera 12, the person detection unit 12c can detect the visitor from the sound detected by the microphone 12b (for example, the voice or footsteps of the visitor).

In the next step S2, when the person detection unit 12c detects a visitor, the feature part detection unit 12d detects the feature part of the visitor from the image of the visitor. In the following description, the case where the feature part detection unit 12 d detects a face that is a feature part of a visitor will be described.

In step S3, the feature quantity extraction unit 12n of the person specifying unit 12e extracts the feature quantity of the face from the image of the visitor's face detected by the feature part detection unit 12d.

In step S4, the similarity determination processing unit 12o of the person specifying unit 12e compares the feature amount of the face extracted by the feature amount extraction unit 12n with the feature amount of the face included in the plurality of pieces of person information stored in the DB 12l. Determine the degree of similarity. Thus, it is possible to specify a person who matches the visitor to be determined.

In step S5, the age / sex estimation processing unit 12p of the person specifying unit 12e makes a visit by comparing the feature amount of the face extracted by the feature amount extracting unit 12n with the plurality of pieces of personal information stored in the DB 12l. Estimate the age and / or gender of the person. Thereby, the age and / or sex of the visitor who was the target of determination can be specified.

By sequentially performing the processes of steps S3 to S5 in this manner, the person specifying unit 12e can specify a visitor.

In the next step S6, the response judging unit 12f receives the identification result of the visitor in the person identifying unit 12e, and determines whether to start a dialogue (conversation) with the visitor.

In this case, when the person identification unit 12e accurately identifies the visitor's person image, the response determination unit 12f determines that interaction with the visitor is possible, and determines to start interaction with the visitor. (Step S6: YES), the process proceeds to the next step S7.

On the other hand, if the person identification unit 12e can not identify the visitor, or if the visitor can be identified but is not accurately identified, the response determination unit 12f determines that the visitor is not identified. It is determined that the dialogue is impossible (step S6: NO), and the process returns to step S1 to perform the processes of steps S1 to S6 again.

In step S7, in response to the decision to execute the dialogue with the visitor in the response judgment unit 12f, the dialogue engine unit 12g selects the most suitable scenario for the dialogue from the DB 12l.

In this case, for example, the related information collection unit 12 h collects peripheral information from the peripheral information DB 22 via the communication unit 12 i and the communication line 14, and collects personal information from the personal information DB 24 and collects the peripheral information and the individual Information is associated with collection time and registered in the DB 12 l as related information including personal information. Therefore, the dialogue engine unit 12g extracts, from the DB 12l, the related information collected at the time when the imaging unit 12a captures an image of the visitor who has determined that interaction with the visitor is possible, and the extracted related information; A scenario corresponding to the determination of the response determination unit 12f is selected from the DB 12l. For example, the dialogue engine unit 12g selects an optimal scenario from the DB 12l in consideration of the installation location of the camera 12 included in the related information and weather information, traffic information and crime prevention information around the installation location.

In step S8, the dialogue engine unit 12g externally outputs the fixed phrase indicated by the scenario as a voice from the speaker 12j. For example, if the visitor is a delivery person in charge of a carrier such as a delivery company, a home delivery company or a postal company, "You have trouble. Please put in the post." And the like, etc. are output from the speaker 12 j. That is, such a scenario is accumulated in the DB 12l, and the dialogue engine unit 12g selects the most suitable scenario for interacting with the business operator in step S7, and selects the most suitable selected scenario in step S8. The sentence is output as sound from the speaker 12 j.

In step S9, the response determination unit 12f determines whether to end the dialog with the visitor. If the dialogue is to be continued, the process returns to step S1, and the processes of steps S1 to S9 are performed again. Thus, the camera 12 can perform moving image shooting or intermittent still image shooting for the visitor by repeatedly executing the processing of steps S1 to S9. Thereby, the characteristic part detection unit 12d detects the visitor's face from the image of the visitor every time the person detection unit 12c detects the visitor in step S2 (every time the process of step S1 is executed). It can detect and track the face. Further, in order to repeatedly execute the processes of steps S1 to S9, in step S6, the response determination unit 12f makes an affirmative determination and continues the execution of the dialog.

Then, in step S9, when the response determination unit 12f determines the end of the dialogue with the visitor (step S9: YES), the process proceeds to the next step S10.

In step S10, the dialogue engine unit 12g records, in the data storage unit 12m, the image (of the visitor) captured by the imaging unit 12a during the dialogue and the sound (voice data of the visitor) detected by the microphone 12b. Further, in step S11, the dialogue engine unit 12g creates a new scenario according to the content of the dialogue with the visitor, and registers the created new scenario in the DB 12l. In this case, the dialogue engine unit 12 g may create a new scenario by correcting or changing the scenario used for the dialogue.

As described above, in the first embodiment, since the processes of steps S1 to S11 are executed each time a visitor visits, the dialogue engine unit 12g performs dialogue learning through dialogue with the visitor, and a new process is performed. Create a scenario and register it in DB12l. As a result, the scenario is accumulated in the DB 12l each time the dialogue is executed, and the scenario in the DB 12l can be updated. As a result, when interacting with a visitor, the dialogue engine unit 12 g can select the latest and optimal scenario from the DB 12 l and use the selected scenario to improve the accuracy of the dialogue.

In step S7, the dialogue engine unit 12g may select an optimal scenario from the latest scenarios (for example, the past one year) registered in the DB 12l. That is, since new scenarios are sequentially registered in the DB 12l as time passes by the above-mentioned dialogue learning, if the latest scenario is selected for the same or similar dialogue, an appropriate dialogue is executed. Is possible.

Note that the image and sound during interaction recorded in the data storage unit 12m in step S10 can be deleted from the data storage unit 12m based on, for example, the input operation of the operation unit 18e by the owner of the portable device 18 is there.

2.2 Second Embodiment
A second embodiment (third correspondence) will be described with reference to FIG. In the second embodiment, the server 16 makes a more detailed identification of the visitor, and the camera 12 interacts with the visitor based on the identification result. Therefore, in the second embodiment, it should be noted that the subject of processing is the camera 12 and the server 16. Also in the second embodiment, as in the first embodiment of FIG. 4, a case where a visitor visits a building from the outside and identifies the visitor based on the face of the visitor will be described.

After the process of step S5 of FIG. 4, the response determination unit 12f of the camera 12 proceeds to step S12 of FIG.

In step S12, the response determination unit 12f determines whether to inquire the server 16 about the detailed identification of the visitor based on the identification result of the person identification unit 12e.

In the case where the person identification unit 12e can specify the person image of the visitor with high accuracy from the image of the visitor's face using the visitor image captured by the imaging unit 12a, the response determination unit 12f It is determined that the inquiry to the server 16 is unnecessary (step S12: NO), and the camera 12 side executes the processing after step S6 in FIG.

On the other hand, even if the image of the visitor can not be accurately identified from the image of the visitor's face even if the image of the visitor taken by the imaging unit 12a is used in the person specifying unit 12e, or If the visitor does not correspond to a plurality of pieces of personal information stored in the DB 12l (if it is a visitor who has visited for the first time), the response judgment unit 12f visits the camera 12 side. It is determined that it is difficult to identify the person, and it is determined to inquire the server 16 about the detailed identification of the visitor (Step S12: YES).

In the next step S13, the response determination unit 12f transmits the inquiry content for requesting detailed identification of the visitor to the server 16 via the communication unit 12i and the communication line 14. In this case, the inquiry contents include instructions for requesting detailed identification of the visitor, determination results by the response determination unit 12f, identification results by the person identification unit 12e, and processing by the person identification unit 12e. The image of the visitor who was (the image of the visitor's face) etc. are included.

When the communication unit 16a of the server 16 receives the inquiry content from the communication unit 12i of the camera 12 via the communication line 14, the communication unit 16a outputs the received inquiry content to the inquiry content judgment unit 16b.

In step S14, the inquiry content determination unit 16b confirms (determines) the inquiry content input from the communication unit 16a. In this case, since the input query content is the query content for requesting the visitor's detailed identification, the query content determination unit 16b outputs the query content to the data matching unit 16c. When the information input from the communication unit 16a is information other than the inquiry content, the inquiry content determination unit 16b does not output the information to the data matching unit 16c.

In step S15, the data matching unit 16c can recognize that the camera 12 has requested the detailed identification of the visitor by inputting the inquiry content from the inquiry content determination unit 16b. Then, after the data matching unit 16c confirms the input inquiry content, the data matching unit 16c acquires the peripheral information around the installation location of the camera 12 from the peripheral information DB 22 via the communication unit 16a, and the visitor from the personal information DB 24 Obtain personal information according to your needs. Next, the data matching unit 16c identifies the visitor in detail by matching the acquired peripheral information and personal information with the contents of the inquiry.

In this case, the data matching unit 16c has the same function as that of the person specifying unit 12e of the camera 12 (of the feature quantity extraction unit 12n, the similarity determination processing unit 12o, and the age / sex estimation processing unit 12p). Therefore, also in the data matching unit 16c, the same process as the person specifying unit 12e is performed using the image or the like of the visitor's face included in the inquiry content. However, in the peripheral information DB 22 and the personal information DB 24, a much larger amount of data is accumulated than the DB 12 l of the camera 12, and information about new visitors that can not be identified by the camera 12 is also accumulated. . Therefore, the data matching unit 16c can identify the visitor in more detail by using the information stored in the DBs 22 and 24.

After the detailed identification of the visitor is performed in this manner, the process proceeds to step S6 in FIG. The response determination unit 16d of the server 16 executes the process of step S6.

That is, in step S6, the response determination unit 16d of the server 16 determines whether to start conversation (conversation) with the visitor based on the detailed identification result for the visitor in the data matching unit 16c. As described above, since the person image of the visitor is specified in detail in the data matching unit 16c, the response determination unit 16d determines that interaction with the visitor is possible, and starts interaction with the visitor. It decides (step S6: YES), it advances to the following step S7.

In step S7, the dialog engine unit 16e of the server 16 receives the decision on execution of the dialog with the visitor in the response determination unit 16d, and selects the optimal scenario for the dialog from the scenario DB 26 via the communication unit 16a, get. In this case, the scenario DB 26 stores much more scenarios than the DB 12 l of the camera 12. Therefore, the dialogue engine unit 16 e can reliably select a scenario optimal for the dialogue with the visitor from among the plurality of scenarios accumulated in the scenario DB 26. Further, in step S7, the dialogue engine unit 16e acquires the peripheral information registered at the time when the image of the visitor was captured from the peripheral information DB 22, and acquires and acquires the personal information of the visitor from the personal information DB 24. It is also possible to select an optimal scenario from the scenario DB 26 in consideration of the surrounding information and personal information.

In step S8, the dialogue engine unit 16e outputs the fixed form sentence indicated by the scenario to the outside as sound from the speaker 12j via the communication unit 16a, the communication line 14 and the communication unit 12i of the camera 12. For example, when the visitor is a person who has visited for the first time, a fixed sentence such as “Who?” Or “What is it for?” Can be output from the speaker 12 j.

In step S9, the response determination unit 16d of the server 16 determines whether to end the dialog with the visitor. When the response determination unit 16d determines the end of the dialog with the visitor (step S9: YES), in the next step S10, the dialog engine unit 16e of the server 16 captures an image of the imaging unit 12a during the dialog (visitor ) And the sound (voice data of the visitor) detected by the microphone 12b are recorded in the history DB 28 as the dialogue history via the communication unit 16a. Further, in step S11, the dialogue engine unit 16e of the server 16 creates a new scenario according to the contents of the dialogue with the visitor using the dialogue history, and registers the created new scenario in the scenario DB 26.

As described above, in the second embodiment, the processes of steps S14 and S15 of FIG. 5 and steps S6 to S11 of FIG. 4 are sequentially executed each time an inquiry content is transmitted from the camera 12 to the server 16. The engine unit 16e performs interaction learning through interaction with the visitor, creates a new scenario using the interaction history of the history DB 28, and registers it in the scenario DB 26. As a result, each time the dialog is executed, a scenario is accumulated in the scenario DB 26, and the scenario in the scenario DB 26 can be updated. As a result, when interacting with a visitor, the dialogue engine unit 16e can select the latest and optimal scenario from the scenario DB 26, and use the selected scenario to improve the accuracy of the dialogue.

In step S7, the dialogue engine unit 16e may select an optimal scenario from the latest scenarios (for example, the past one year) registered in the scenario DB 26. That is, since new scenarios are sequentially registered in the scenario DB 26 with the passage of time by the above-described dialogue learning, if the latest scenario is selected for the same or similar dialogue, the appropriate dialogue is executed. It is because it becomes possible.

In the second embodiment, when the data matching unit 16c can not specify the visitor in detail in step S6 of FIG. 4, the response determination unit 16d of the server 16 can not interact with the visitor. It is determined that (step S6: NO), the process returns to step S1, and the processes of steps S1 to S5 and S12 to S15 are performed again. When the response determination unit 16d of the server 16 determines to continue the dialogue with the visitor in step S9 in FIG. 4 (step S9: NO), the process returns to step S1, and steps S1 to S5, and S12 to S15. , S6 to S9 are sequentially executed again.

Furthermore, in the above description of the second embodiment, the case where all the processes of steps S6, S7, and S9 to S11 of FIG. 4 are performed on the server 16 side has been described. In the second embodiment, part or all of the processing in these steps may be performed on the camera 12 side.

For example, the process of steps S14 and S15 of FIG. 5 is performed on the server 16 side, and in step S6 of FIG. 4 from the response determination unit 16d of the server 16 to the communication unit 16a, the communication line 14, and the communication unit 12i of the camera 12. The identification result of the data matching unit 16c may be transmitted to the response determination unit 12f, and the process of step S6 may be performed by the response determination unit 12f of the camera 12. In this case, the processes after step S7 are executed on the camera 12 side as in the first embodiment.

Alternatively, the processes of steps S6, S7, and S9 in FIG. 4 are performed on the camera 12 side, and the dialog engine unit 16e of the server 16 receives the positive determination result of step S9 by the response determination unit 12f on the camera 12 side. The interaction history may be recorded in the history DB 28 in step S10, and a new scenario may be created using the interaction history in step S11 and accumulated in the scenario DB 26.

<2.3 3rd Example>
A third embodiment (second correspondence) will be described with reference to FIG. In the third embodiment, the camera 12 notifies the portable device 18 of the visit of the visitor, and then the dialogue between the visitor and the holder of the portable device 18 is performed. Therefore, it should be noted that in the third embodiment, the subject of processing is the camera 12 and the portable device 18. Also in the third embodiment, as in the first embodiment shown in FIG. 4, a case where a visitor visits a building from the outside and identifies the visitor based on the face of the visitor will be described.

After the process of step S8 in FIG. 4, the response determination unit 12f of the camera 12 proceeds to step S16 in FIG.

In step S16, the response determination unit 12f determines, based on the identification result of the visitor in the person identification unit 12e, whether to notify the portable device 18 of the visit of the visitor.

When it is determined that the notification of the visit is unnecessary (step S16: NO), the response determination unit 12f, because the camera 12 can cope with the visit, such as the visit of a company such as a delivery company, a home delivery company, or a postal company. Executes the process of step S9 of FIG.

On the other hand, when it is determined that the mobile device 18 needs to be notified of the visit of the visitor, such as the case where the visitor wants to have a conversation with the resident of the house, or the case of the visitor visiting for the first time (step S16: YES) , And proceeds to the next step S17.

In step S17, the response determination unit 12f checks the portable device 18 via the communication unit 12i and the communication line 14 for notification of the visitor's visit and the presence or absence of start of dialog (TV telephone service) with the visitor. Send the inquiry content for.

In step S18, when the communication unit 18a of the portable device 18 receives the inquiry content from the camera 12, the display unit 18d displays the inquiry content, while the speaker 18f outputs a sound indicating the inquiry content. The owner of the portable device 18 looks at the display content of the display unit 18 d or listens to the sound from the speaker 18 f to make an inquiry about whether the visitor is visiting or the start of the video telephone service is the camera 12. It can be recognized that it was from.

In step S19, the holder of the portable device 18 determines whether to start the video telephone service based on the display content of the display unit 18d and / or the sound from the speaker 18f.

When the video telephone service is not started, that is, when the dialog with the visitor is not performed (step S19: NO), the holder of the portable device 18 does not operate the operation unit 18e at all. As a result, no reply is sent from the portable device 18 to the camera 12 via the communication line 14. Therefore, the response determination unit 12f of the camera 12 may receive a predetermined time from the time when the content of the inquiry is transmitted in step S17. If there is no response from the portable device 18, it is determined that the holder does not have an intention to start the video telephone service, and the process proceeds to step S9 in FIG.

Alternatively, when the video telephone service is not started (step S19: NO), the owner of the portable device 18 operates the operation unit 18e and inputs an instruction content indicating that the video telephone service is not started. Thereby, the communication unit 18 a transmits an instruction signal corresponding to the input instruction content to the communication unit 12 i of the camera 12 via the communication line 14. The response determination unit 12f of the camera 12 recognizes that there is an instruction not to start the video telephone service based on the instruction signal received by the communication unit 12i, and proceeds to the process of step S9 in FIG.

On the other hand, when it is desired to start the video telephone service (step S19: YES), the holder of the portable device 18 operates the operation unit 18e and instructs the start of the video telephone service. Thereby, the communication unit 18 a transmits an instruction signal corresponding to the input instruction content to the communication unit 12 i of the camera 12 via the communication line 14.

In step S20, the response determination unit 12f of the camera 12 recognizes that the start of the video telephone service has been instructed based on the instruction signal received by the communication unit 12i, and determines the start of the video telephone service.

The dialogue engine unit 12g receives the determination by the response determination unit 12f, and (the communication unit 18a of) the portable device 18 via the communication unit 12i and the communication line 14, the image of the visitor captured by the imaging unit 12a, The voice data of the visitor detected by the microphone 12b is transmitted.

The display unit 18d of the portable device 18 displays the image of the visitor received by the communication unit 18a, and the speaker 18f outputs the sound data received by the communication unit 18a as a sound. The holder of the portable device 18 visually recognizes the image of the visitor displayed on the display unit 18 d and after hearing the sound from the speaker 18 f, starts a conversation with the visitor. The microphone 18g detects the voice of the holder, and the communication unit 18a transmits the voice (voice data) of the holder detected by the microphone 18g to the communication unit 12i of the camera 12 via the communication line 14.

The speaker 12j of the camera 12 outputs the sound data of the possessor received by the communication unit 12i as a sound. Thus, the visitor can talk with the owner (resident of the house) by listening to the sound from the speaker 12 j. In this way, a dialogue (TV phone service) is started between the visitor and the holder of the portable device 18.

In step S21, the response determination unit 12f determines whether to end the video telephone service. In this case, the response determination unit 12 f receives an instruction signal instructing termination of the video telephone service from the portable device 18 via the communication line 14 by the communication unit 12 i by the operation of the operation unit 18 e by the owner of the portable device 18. Check if it has been done.

When the instruction signal is not received (step S21: NO), the response determination unit 12f determines the continuation of the video telephone service. In response to the determination result of the response determination unit 12f, the dialog engine unit 12g executes the process of step S20 and continues the video telephone service.

On the other hand, when the instruction signal instructing the end of the video telephone service is received by the communication unit 12i in step S21 (step S21: YES), the response judging unit 12f determines the end of the video telephone service. The dialogue engine unit 12 g stops the transmission of the image and voice data of the visitor to the portable device 18 through the communication unit 12 i and the communication line 14 in response to the determination of the response determination unit 12 f. On the other hand, also in the portable device 18, the transmission of the voice data of the owner to the camera 12 is stopped along with the generation of the instruction signal instructing the end of the video telephone service. As a result, the video telephone service ends.

In step S22, the dialog engine unit 12g records in the data storage unit 12m the image of the visitor captured by the imaging unit 12a during the dialogue (during the video telephone service) and the voice data of the visitor detected by the microphone 12b. . Also in this case, for example, based on the input operation of the operation unit 18e by the owner of the portable device 18, it is possible to delete the recorded image and sound during the interaction from the data storage unit 12m.

<2.4 Modifications of the First to Third Embodiments>
Next, modified examples (first to fourth modified examples) of the first to third embodiments will be described. In the description of the first to fourth modifications, the operations of the first to third embodiments described above are applied to the operations not particularly described.

(2.4.1 First Modification)
The first modified example is a response when a resident of a house enters and leaves. For example, when the owner of the portable device 18 goes out and is absent while the other resident is in the house, when the other resident goes out of the house, as described below, The modification is applied to a part of the operation of the third embodiment.

Specifically, when applied to the first embodiment, the imaging unit 12a captures an image of another resident going out of the house in step S1 of FIG. 4, and the person detecting unit 12c captures another resident based on the captured image. To detect In the next step S2, the feature part detection unit 12d detects a feature part of another resident, for example, the face or appearance (back view) of another resident.

Then, in steps S3 to S5, the person specifying unit 12e specifies another resident from the feature portion (face or back view). After that, the process after step S6 is performed, and by performing the dialog with other residents, the camera 12 can perform a call to the other residents who are going out of the house. .

Alternatively, in the first modification, after the processing of step S8, the third embodiment of FIG. 6 may be executed, or as shown by a broken line in FIG. 4, the processing of steps S6 to S8 may be omitted. The third embodiment of FIG. 6 may be implemented. In this case, in step S16 in FIG. 6, the response determination unit 12f determines to notify (inform) the portable device 18 that another resident is going out of the house (step S16: YES). In the next step S17, the response determination unit 12f notifies the communication unit 18a of the portable device 18 that the other resident is going out of the house through the communication unit 12i and the communication line 14. Thereby, in step S18, the display unit 18d of the portable device 18 displays the notification content, and the speaker 18f outputs a sound according to the notification content.

The holder of the portable device 18 recognizes that the other resident will go out of the house by visually recognizing the display content of the display unit 18 d or by listening to the sound from the speaker 18 f. As described above, when another resident goes out of the house, the portable device 18 is always notified, so the owner of the portable device 18 executes a dialogue with the other resident in steps S19 and S20. Etc, you can take necessary measures.

In the first modification, if it is difficult to specify another resident in the process of steps S3 to S5 in FIG. 4, the second embodiment is applied, and the other resident is specified in detail on the server 16 side. It is also good. Also, in the above description, the back part of the other resident is the characteristic part, but the face of the other resident, the figure of the other resident viewed from the front side, the side or side, the clothes of the other resident, etc. It may be Since these characteristic parts are registered in the DB 12l as personal information (including related information), in the person specifying part 12e, images and characteristic quantities of characteristic parts of other residents and personal information (characteristic parts in It is possible to identify other residents in contrast to information).

(2.4.2 second modification)
The second modified example changes the operation of a part of the first and second embodiments, and adds an image of the visitor's face, etc. or an image of the visitor's clothes etc. instead of the image of the visitor's face To identify the visitor.

In this case, in the first embodiment, in step S2 of FIG. 4, the feature detection unit 12d detects the face or appearance that is the feature of the visitor from the image of the visitor detected by the person detection unit 12c. Furthermore, the clothes of the visitor, the pattern, the color or the pattern of the clothes, or the marks attached to the clothes of the visitor are detected.

In step S3, the feature quantity extraction unit 12n of the person specifying unit 12e extracts the feature quantity of the face or appearance from the image of the face or appearance of the visitor detected by the feature part detection unit 12d. The feature quantities of these features are extracted from the image of the clothes, the pattern of the clothes, the color or pattern, or the mark attached to the clothes of the visitor.

In step S4, the similarity determination processing unit 12o compares the feature amount of each feature portion extracted by the feature amount extraction unit 12n with the feature amount of the feature portion included in the plurality of pieces of personal information stored in the DB 12l. Determine Further, in step S5, the age / sex estimation processing unit 12p compares the feature amount of each feature extracted by the feature amount extracting unit 12n with the plurality of pieces of personal information stored in the DB 12l. Estimate the age and / or gender of

Thus, in addition to the face or appearance of the visitor, the person specifying unit 12e can identify the visitor based on the clothes of the visitor, the pattern, color or pattern of the clothes, or the mark attached to the clothes. It can be identified.

Thereafter, by executing the processes of steps S6 to S9, the camera 12 executes the dialogue with the visitor. As a case where such processing is performed, for example, when the visitor is a business (for example, a delivery company, a delivery company, or a delivery person in charge of a mail business), the camera 12 is installed when the business visits When there is no one in the house where you are working, it means that only the resident who has difficulty in dealing with the employer is in the house, or the resident does not want to attend the employer. In addition, the resident who has difficulty in dealing with a business person means, for example, a child who is difficult to handle a business person, or a person who is being treated or recuperated in a house because of illness or the like.

Therefore, in step S7, the dialogue engine unit 12g selects a scenario optimal for dialogue in such a case from the DB 12l.

Specifically, related information corresponding to the above case is registered in the DB 12 l as related information of the installation place (house) of the camera 12. Therefore, the dialogue engine unit 12 g extracts the relevant information from the DB 12 l, and selects a scenario corresponding to the extracted relevant information and the determination content in the response determination unit 12 f from the DB 12 l.

Thus, in step S8, the dialogue engine unit 12g outputs the fixed form sentence indicated by the selected scenario to the outside as a voice from the speaker 12j. For example, a fixed phrase such as “Please put in a delivery locker”, “I would like to re-deliver at the latest time slot”, etc. is output as an audio from the speaker 12 j. As a result, if there is no one in the house, or if only the resident who has difficulty in responding to the employer is in the house, or if the resident does not want to attend the employer, the camera 12 It is possible to take appropriate measures for visitors.

In the case of the second embodiment, in the data matching process of step S15 of FIG. 5, the data matching unit 16c adds to the face or appearance of the visitor as in the process by the person specifying unit 12e described above. Since the visitor may be specified based on the clothes of the visitor, the pattern, color or pattern of the clothes, or the marks attached to the clothes, the detailed description will be omitted.

(2.4.3 third modification)
The third modification is a modification in which the information on the visitor is registered in advance in the DB 12l when the visitor's visit is known.

In step S31 of FIG. 7, for example, the visitor is an enterprise such as a delivery company, a home delivery company, or a postal company, and from the enterprise to the holder of the portable device 18 or the portable device 18, the residence of the holder. It is assumed that the scheduled visit time to the site (scheduled delivery time) has been notified in advance (step S31: YES). Note that, for example, if the mobile device 18 is a smartphone, this notification is performed by e-mail transmission from the business entity to the mobile device 18, a telephone call, or a social network service (SNS).

In step S32, when the owner leaves the house at the scheduled visit time, the owner operates the operation unit 18e of the portable device 18, inputs information on the business operator who is the visitor, and the camera at the scheduled visit time 12 instructs to take a predetermined response. In this case, as the information input by the owner, for example, the information of the visitor (the name of the company, the name of the person in charge of the company (visitor), etc.), the purpose of the visit, and the purpose of the visit are delivery. , Information on the item to be delivered (the name of the sender, etc.), scheduled visit time, etc.

In step S33, the communication unit 18a of the portable device 18 transmits the input information to the communication unit 12i of the camera 12 via the communication line 14. Thus, in step S34, the related information collection unit 12h registers the information received by the communication unit 12i in the DB 12l as related information (and personal information).

Thereby, in the first embodiment of FIG. 4, in the camera 12, when the company visits a house at the scheduled visit time, the processes after step S 1 are executed. Specifically, whether or not the visitor who has visited is a business is specified by the processing of steps S1 to S5. Thereafter, in the process of step S7, an optimal scenario corresponding to the related information (information input by the owner) registered in advance in the DB 12l is selected from the DB 12l. As a result, even if the owner of the portable device 18 is away from home in step S8, the dialog engine unit 12g can execute the dialog with the visiting business operator according to the selected scenario.

In addition, in step S31 of FIG. 7, when a visitor's visits, such as a company etc. do not understand beforehand (step S31: NO), since the owner of the portable apparatus 18 does not input said information, the camera 12 Executes the processing of the first embodiment of FIG.

In addition, if the company visits earlier than the scheduled visit time due to traffic conditions around the house, or if the provider visits later than the scheduled visit time, the actual visit time matches the scheduled visit time Except for the point where it does not, the business operator who has visited and the business operator indicated by the related information (person information) match. In this case, the response determination unit 12f determines the execution of the dialog with the business operator who visited in step S6 of FIG. 4, and the dialog engine unit 12g selects the corresponding scenario from the DB 12l based on the related information in step S7. do it. Even in this case, the dialog engine unit 12g can execute the dialog with the visiting business operator according to the selected scenario.

On the other hand, when a visitor (for example, another company) other than the company originally planned to visit visits the scheduled time, etc., the above-mentioned first embodiment is applied, and the response judgment unit 12f performs the step The execution of the dialog with the visitor is determined in S6, and the dialog engine unit 12g may select the corresponding scenario from the DB 12l in step S7. Thereby, even if the scene is different from the original schedule, the dialogue engine unit 12 g selects an appropriate scenario according to the scene, and executes the dialogue with the visitor according to the selected scenario. it can.

Alternatively, when a visitor other than the company originally planned to visit visits at the scheduled arrival time etc., the third embodiment described above is applied, and the response determination unit 12f sends the portable device 18 in step S16 of FIG. The notification may be determined (step S16: YES), and the communication unit 12i may notify the communication unit 18a of the portable device 18 via the communication line 14 that different visitors have visited. Thereby, the display unit 18d of the portable device 18 displays the notification content received by the communication unit 18a, and the speaker 18f outputs a sound according to the notification content to the outside. As a result, the holder of the portable device 18 can understand from the display content of the display unit 18 d or the sound of the speaker 18 f that a visitor different from the original schedule has visited the house.

Furthermore, in the third modification, when the enterprise visits at the scheduled visit time in combination with the second modification, when there is no one in the house, when only residents who have difficulty in dealing with the enterprise are in the house, or When the resident does not want to respond to the business, it is possible to respond as described above. In this case, the camera 12 executes a dialog with the business entity, and can take, for example, a response to change the visit date and time, etc., because it is absent.

As described above, in the third modification, which business entity delivers the package from whom (sender) from the portable device 18 to the camera 12 (visit scheduled time, designated date and time, and time zone) When the information indicating the information is transmitted and the transmitted information is used to specify the visitor, it is possible to determine on the camera 12 side whether or not the person is an authentic business person indicated in the information.

(2.4.4 fourth modification)
In the fourth modification, even if the visitor 12 is identified by the camera 12, the visitor can not be identified with high accuracy, or even if the dialogue with the visitor is performed, the visitor does not It corresponds to the case where there is no reaction. The fourth modified example assumes that the visitor is a suspicious person.

In this case, if the visitor is a suspicious person, the person information of the suspicious person registered in the DB 12l is not registered, so even if the processing of steps S1 to S5 in FIG. It can not be identified. Therefore, the response determination unit 12f executes the processes of steps S16 and S17 of FIG. 6, determines that the visitor is a suspicious person, and determines the determination result to the portable device 18 through the communication unit 12i and the communication line 14. To the communication unit 18a. In step S18, the display unit 18d of the portable device 18 displays the notification received by the communication unit 18a, and the speaker 12j outputs a sound according to the content of the notification. The owner of the portable device 18 understands that the suspicious person has visited by looking at the display contents of the display section 18d or by listening to the sound from the speaker 12j, and the dialogue with the suspicious person after step S19, etc. You can take the necessary action.

Alternatively, if the visitor can not be identified with high accuracy even if the processing in steps S1 to S5 of FIG. 4 is performed, the response determination unit 12f determines that the visitor is a suspicious person in step S6, and It decides to interact with the suspicious person (step S6: YES). Then, the dialogue engine unit 12g selects a scenario in step S7, and executes dialogue in step S8. In this case, even if the suspicious person calls from the camera 12, the suspicious person may not respond at all. In such a case, the processes after step S16 in FIG. 6 are executed as in the above-described correspondence, and the camera 12 notifies the portable device 18 of the visit of a suspicious person. As a result, the owner of the portable device 18 can take necessary measures such as dialogue with a suspicious person.

In any of the above two examples, in the camera 12, the imaging unit 12a images a suspicious person, the microphone 12b detects the voice of the suspicious person, and the captured image of the suspicious person and data of the suspicious person's voice are It is desirable to record in the data storage unit 12m together with the information on the time of imaging. In the above description, although the visitor is a suspicious person, the above processing is also performed when the response determination unit 12 f estimates that the visitor is a suspicious person, even if the visitor is not actually a suspicious person. May be performed.

(2.4.5 Other Modifications)
In the above description, when the visitor is identified from the image of the visitor's face mainly for the visitor who visits the building from the outside (the first to third embodiments etc.) or outside the house The case where the other resident is specified from the image of the back view of the other resident (the first modified example) has been described for the other resident who has left.

In this embodiment, the first to third examples and the first to fourth modified examples are combined, and in addition to or in place of the face of the visitor, the visit is performed from other characteristic portions (such as appearance and clothes). It is also possible to identify the Also, with regard to other residents going out of the house, in addition to or in place of the back view, the other features are identified from other features (face, appearance from other directions, clothes, etc.) It is also possible. In this case, the second embodiment can be applied to the identification of the visitor or other resident instead of the identification on the camera 12 side, and the identification on the server 16 side is also possible.

[3. Effect of this embodiment]
As described above, according to the security system 10 according to the present embodiment, the artificial intelligence is mounted on the camera 12, and the visitor 12 is made to function as the person detection unit 12c, the characteristic part detection unit 12d, and the person identification unit 12e. It is possible to provide a service tailored to a specific person, such as, at the camera 12 side as a local area. Thereby, for example, even if the resident (owner of the portable device 18) of the house where the camera 12 is installed is absent, the camera 12 can take an appropriate response to the visitor. As a result, the convenience of the security system 10 can be improved.

In this case, the person detection unit 12c detects, for the image captured by the camera 12, a moving object reflected in the image, a change in the noise of the image, or a change in the brightness of the image. To detect a person. Thus, a person can be easily detected using only the image captured by the imaging unit 12a. As a result, with the detection result of the person detection unit 12c as a trigger, the characteristic part detection unit 12d and the person specification unit 12e can execute each process.

Further, the person detection unit 12c can detect a person with high accuracy by using the image captured by the imaging unit 12a and the sound detected by the microphone 12b in combination.

Furthermore, when the camera 12 images the surroundings of the camera 12 continuously or at predetermined time intervals, the characteristic part detection unit 12 d detects the characteristics of the person from the image of the person every time the person detection unit 12 c detects the person. By detecting the part and tracking the feature part, it is possible to easily grasp the behavior of a person such as a visiting visitor, and to take an appropriate response to the person.

The person specifying unit 12e extracts the feature amount of the feature portion from the image of the feature portion of the person detected by the feature portion detecting unit 12d, and the feature of the person extracted by the feature amount extracting unit 12n. Similarity determination processing unit 12o that determines the similarity between the feature amount of the portion and the feature amount of the feature portion of the person indicated by the person information, the feature amount of the feature portion of the person extracted by the feature amount extraction unit 12n, and the person information Since the age / sex estimation processing unit 12p estimates the age and / or sex of the person detected by the person detection unit 12c by comparing with each other, the person detected by the person detection unit 12c is any person. It becomes possible to identify easily and efficiently.

In this case, if at least the face or appearance of the person detected by the person detection unit 12c is a characteristic part, when the camera 12 is installed in a building such as a house, a visitor to the building or entering or leaving the building It is possible to easily detect the movement of the resident.

In addition, when the camera 12 is installed in a building such as a house by including the clothes of a person, the pattern, color or pattern of the clothes, or the mark attached to the clothes in the characteristic part, It is possible to more easily identify what kind of person the visitor of or the person entering or leaving the building is.

Furthermore, when the person specifying unit 12e specifies a person, the camera 12 has a response determining unit 12f that determines the response to the person according to the specified result, so that an appropriate response according to the person image of the person is made. It is possible to take

In the security system 10, by applying the first embodiment, the camera 12 can easily interact with a person according to a scenario. In addition, since the camera 12 has the dialogue engine unit 12g which is a function of artificial intelligence, even if the amount of image and voice data increases due to the execution of dialogue, the dialogue engine unit 12g easily and quickly It becomes possible to process. Furthermore, since the dialogue engine unit 12 g performs dialogue learning through dialogue with a person, it is possible to improve various functions of artificial intelligence including the dialogue function.

In addition, when the response determination unit 12f determines execution of a dialog with a person, the dialogue engine unit 12g collects the related information collection unit 12h at the time when the imaging unit 12a captures an image of the person specified by the person specification unit 12e. The related information is extracted from the DB 12 l, and a scenario corresponding to the extracted related information and interaction is selected from the DB 12 l. Therefore, referring to the related information, a more appropriate scenario is selected, and the accuracy of the interaction with the person is determined. It can be improved. In addition, when the camera 12 is installed in a building such as a house, if a visitor to the building is known in advance, the visitor visited the building if the relevant information of the visitor is registered in the DB 12 l In doing so, appropriate dialogue can be performed for the visitor.

Furthermore, the camera 12 includes the related information collection unit 12h, whereby the latest related information is registered in the DB 12l, so that it is possible to further improve the accuracy of the dialog with the person.

In addition, when the response determination unit 12f determines execution of the dialog with the specified person by the person specifying unit 12e, the dialog engine unit 12g collects the related information collection unit 12h at the time when the imaging unit 12a captures a person. Since the related information is extracted from the DB 12l and the extracted related information and the scenario corresponding to the interaction are selected from the DB 12l, the accuracy of the interaction with the person is further improved.

Furthermore, when the related information according to the person specified by the person specifying unit 12e is not registered in the DB 12l, the dialogue engine unit 12g extracts other related information registered in the DB 12l and extracts the other related information The scenario corresponding to the information and the dialogue is selected from the DB 12 l and / or the portable device 18 is informed that there is no corresponding person. Thereby, when the camera 12 is installed in a building such as a house, even when the visitor to the building and the visitor indicated by the related information registered in the DB 12l do not match, other related information is It can be used to carry out a dialogue with the visitor. In addition, it is possible to notify the holder (resident) of the portable device 18 who is away from the building of the visit of the visitor.

In addition, when the security system 10 applies the second embodiment, the peripheral information DB 22 and the personal information DB 24 are provided on the server 16 side, and the amount of data necessary to specify a person is determined by the camera 12. There are also many cases, and by matching the peripheral information and personal information with the contents of the inquiry from the camera 12, the server 16 can specify the person's image etc. in detail. In addition, by installing artificial intelligence in the server 16 and making it function as the inquiry content determination unit 16b and the data matching unit 16c, it is possible to provide a service tailored to a person on the server 16 side as a local area. . Even in this case, even if there is no resident of the house where the camera 12 is installed, the server 16 can take an appropriate response to the person. As a result, the convenience of the security system 10 can be improved. Furthermore, by matching data, it is possible to easily associate a person with content such as peripheral information and personal information.

In addition, since the server 16 includes the response determination unit 16d and the dialogue engine unit 16e, the camera 12 can easily and accurately interact with a person according to the scenario selected on the server 16 side. Further, since the server 16 has the dialogue engine unit 16e which is a function of artificial intelligence, even if the amount of image and voice data increases by execution of the dialogue, the dialogue engine unit 16e can easily and quickly It is possible to process Furthermore, the dialogue engine unit 16e can perform dialogue learning through dialogue with a person and improve various functions of artificial intelligence including the dialogue function.

In addition, after the end of the dialogue, the dialogue engine unit 16e creates a new scenario using the dialogue history recorded in the history DB 28, and registers the created new scenario in the scenario DB 26, so that the dialogue engine unit 16e It is possible to confirm at a later date what kind of interaction has been performed, and the convenience of the security system 10 can be further improved. Further, since artificial intelligence is installed in the server 16, even if the amount of data accumulated in the history DB 28 dramatically increases, the dialogue engine unit 16e processes the data appropriately and promptly and is created. New scenarios can be accumulated. Moreover, since a new scenario is created using the dialogue history accumulated in the history DB 28, the dialogue engine unit 16e can easily select an optimal and latest scenario when interacting with a person. Become. That is, since the dialogue engine unit 16e creates a new scenario by dialogue learning through dialogue with a person, it is easy to select an optimal scenario each time the dialogue with a person is executed.

Furthermore, in the security system 10, by applying the third embodiment, the holder of the portable device 18 who is out of the installation location of the camera 12 can be installed in the installation location of the camera 12, even if not at the installation location. Notification of the presence of a person (eg, notification of a visitor's visit) can be received in real time. Thus, the security system 10 can provide the holder with appropriate services.

In the third embodiment, after (the holder of) the portable device 18 receives the notification, it is possible to interact with the person, and the person on the camera 12 side and the holder of the portable device 18 By starting the dialogue between them, the holder can receive an appropriate service (dialogue service) in real time without being at the installation place of the camera 12. Thereby, the convenience of the security system 10 can be further improved. Even in this case, the dialogue engine unit 12 g can perform dialogue learning through dialogue with a person and improve various functions of artificial intelligence including the dialogue function.

In the first and third embodiments, when the dialog with the person ends, the camera 12 at least includes the image of the person captured by the imaging unit 12a during the dialog and the voice of the person detected by the microphone 12b. The data is stored in the data storage unit 12m, and the dialogue engine unit 12g creates a new scenario based on the contents of the dialogue with the person after the dialogue ends, and registers the created new scenario in the DB 12l.

As a result, it is possible to confirm at a later date what kind of dialogue has been performed with the person, and the convenience of the security system 10 can be further improved. Further, since the artificial intelligence is mounted on the camera 12, even if the amount of data stored in the data storage unit 12m dramatically increases, the dialogue engine unit 12g processes the data appropriately and promptly, Scenarios can be created and registered in DB 12l. Furthermore, since the dialogue engine unit 12 g creates a new scenario by dialogue learning through dialogue with a person, it is easy to select an optimal scenario each time the dialogue with a person is executed.

Thus, since the artificial intelligence function in the camera 12 and the server 16 is improved by the dialog learning, for example, when it is confirmed that the person is a resident of the house in cooperation with the home security system of the house, It is also possible to take measures such as releasing the locked state of the door.

The present invention is not limited to the above-described embodiment, and it goes without saying that various configurations can be adopted without departing from the scope of the present invention.

Claims (19)

In a security system (10) having at least a camera (12) for imaging a person,
The camera (12) is
An imaging unit (12a) for imaging the surroundings of the camera (12);
A person detection unit (12c) for detecting the person based on an image captured by the imaging unit (12a);
A feature portion detection unit (12d) for detecting a feature portion of the person from the image of the person detected by the person detection unit (12c);
A person identification unit (12e) for identifying the person based on the image of the feature portion detected by the feature portion detection unit (12d);
A security system (10) characterized by having: The security system (10) according to claim 1, wherein
The person detection unit (12c) detects a moving object reflected in the image, detects a change in noise of the image, or detects the brightness of the image in the image captured by the camera (12). A security system (10) characterized in that the person is detected by detecting a change in height. The security system (10) according to claim 1 or 2,
The camera (12) further comprises a first microphone (12b) for detecting sound around the camera (12),
The security is characterized in that the person detection unit (12c) detects the person based on the image captured by the imaging unit (12a) and the sound detected by the first microphone (12b). System (10). In the security system (10) according to any one of claims 1 to 3,
When the camera (12) images the surroundings of the camera (12) continuously or at predetermined time intervals, the characteristic part detection unit (12d) detects the person by the person detection unit (12c). A security system (10) characterized by tracking the feature by detecting the feature from an image of the person. The security system (10) according to any one of claims 1 to 4,
The camera (12) further includes a person information registration unit (12 l) in which person information on a predetermined person is registered.
The person identification unit (12e)
A feature amount extraction unit (12n) for extracting feature amounts of the feature portion from the image of the feature portion detected by the feature portion detection unit (12d);
A similarity determination processing unit (12o) that determines the similarity between the feature amount of the feature portion of the person extracted by the feature amount extraction unit (12n) and the feature amount of the feature portion of the person indicated by the person information;
The age and / or gender of the person detected by the person detection unit (12c) is estimated by comparing the feature information of the feature portion of the person extracted by the feature amount extraction unit (12n) with the person information Age and gender estimation processor (12p),
Security system (10) characterized by comprising. The security system (10) according to any one of claims 1 to 5,
The security system (10), wherein the characteristic part is at least a face or appearance of a person detected by the person detection unit (12c). In the security system (10) according to claim 6,
A security system (10) characterized in that the characteristic portion includes clothes of the person, a pattern, color or pattern of the clothes, or a mark attached to the clothes. A security system (10) according to any one of the preceding claims,
The camera (12) further includes a first response determination unit (12f) that determines a response to the person according to the specification result when the person specification unit (12e) specifies the person. A security system characterized by (10). The security system (10) according to claim 8, wherein
The camera (12) is
Scenario registration unit (12 l) where multiple scenarios are registered,
A first dialogue engine that selects a scenario corresponding to the dialogue from the scenario registration unit (12l) when the first response judging unit (12f) decides to execute dialogue with the person based on the identification result Department (12g),
A first speaker (12j) outputting the scenario selected by the first dialogue engine unit (12g) to the outside as voice;
A second microphone (12b) for detecting the voice of the person;
A security system (10). In the security system (10) according to claim 9,
The camera (12) is a related information registration unit (12l) in which information on the installation place of the camera (12), information on the periphery of the installation place, and / or information on a predetermined person is registered as related information. In addition,
The first dialogue engine unit (12g), when the first response judging unit (12f) decides to execute the dialogue with the person specified by the person specifying unit (12e) based on the specifying result, A security system (10) characterized by extracting the related information from a related information registration unit (12l) and selecting a scenario corresponding to the extracted related information and the dialog from the scenario registration unit (12l). The security system (10) according to claim 10,
The camera (12) further includes a related information collection unit (12h) that collects the related information from the outside and registers the collected related information in the related information registration unit (12l). (10). In the security system (10) according to claim 11,
The first dialogue engine unit (12g), when the first response judging unit (12f) decides to execute the dialogue with the person specified by the person specifying unit (12e) based on the specifying result, The related information collecting unit (12h) extracts the related information collected by the related information collecting unit (12h) when the image is taken by the imaging unit (12a) from the related information registration unit (12l), and the extracted related information and the dialog A security system (10) characterized in that the selected scenario is selected from the scenario registration unit (12l). The security system (10) according to any one of claims 10 to 12,
The first dialogue engine unit (12g) does not register related information according to the person specified by the person specifying unit (12e) in the related information registration unit (12l), the related information registration unit (12g) 12l) extract the other related information registered in 12l), select the extracted other related information and the scenario corresponding to the interaction from the scenario registration unit (12l), and / or notify the outside Security system (10) characterized in that. The security system (10) according to any one of claims 9 to 13,
It further comprises a portable device (18) connectable to the camera (12) via a communication line (14),
The camera (12) further includes a first communication unit (12i) that can communicate with at least the portable device (18) via the communication line (14).
The first communication unit (12i) is configured to communicate with the communication line when the first response judging unit (12f) decides to notify the portable device (18) of the presence of the person based on the identification result. (14) A security system (10) characterized by notifying the portable device (18) of the presence of the person via (14). The security system (10) according to claim 14, wherein
The portable device (18) includes a second communication unit (18a) capable of communicating with at least the camera (12) via the communication line (14), a display unit (18d), and an operation unit. (18e), a second speaker (18f), and a third microphone (18g),
When the second communication unit (18a) receives the notification of the presence of the person, the display unit (18d) displays the notification and / or the second speaker (18f) sounds the notification Output as
When the holder of the portable device (18) operates the operation unit (18e) and instructs start of dialogue with the person, the second communication unit (18a) transmits the communication line (14). Sending an instruction signal instructing the start of the dialogue to the first communication unit (12i);
The first communication unit (12i) receives the instruction signal, the image of the person captured by the imaging unit (12a), and the voice data of the person detected by the second microphone (12b) Are transmitted to the second communication unit (18a) via the communication line (14),
The display unit (18d) displays the image of the person received by the second communication unit (18a), and the second speaker (18f) receives the image of the person received by the second communication unit (18a). Voice data is output as sound, and the second communication unit (18a) transmits the voice data of the holder detected by the third microphone (18g) to the first voice data via the communication line (14). Send to the communication unit (12i),
A security system (10) characterized in that the first speaker (12j) outputs the sound data of the possessor received by the first communication unit (12i) as a sound. The security system (10) according to any one of claims 9 to 13, 15
After the end of the dialog with the person, the camera (12) is at least an image of the person captured by the imaging unit (12a) during the dialog and an image of the person detected by the second microphone (12b). It further has a data storage unit (12 m) in which audio data is recorded,
After completion of the dialogue, the first dialogue engine unit (12g) creates a new scenario based on the contents of the dialogue with the person, and registers the created new scenario in the scenario registration unit (12l). A security system (10) characterized in that. The security system (10) according to claim 8, wherein
A server (16) connectable to the camera (12) via a communication line (14);
A peripheral information database (22) connected to the server (16) and registered with peripheral information including information on the installation location of the camera (12) and information on the periphery of the installation location;
A personal information database (24) connected to the server (16) and in which personal information of a predetermined person is registered;
And have
When it is determined that the first response judging unit (12f) inquires the server (16) about the detailed identification of the person detected by the person detecting unit (12c) based on the identification result, the communication line (14) Query the server (16) for further identification of the person,
The server (16) is
A third communication unit (16a) that receives an inquiry content from the first response determination unit (12f);
An inquiry content judgment unit (16b) for judging the contents of the inquiry received by the third communication unit (16a);
The peripheral information is acquired from the peripheral information database (22) via the third communication unit (16a) based on the determination result of the inquiry content judgment unit (16b), and from the personal information database (24) A data matching unit (16c) for specifying the person in detail by acquiring the personal information according to the person indicated by the contents of the inquiry and collating the acquired peripheral information and the personal information with the contents of the inquiry ,
A security system (10) characterized by having: In the security system (10) according to claim 17,
It further has a scenario database (26) connected to the server (16) and registered with a plurality of scenarios,
When the data matching unit (16c) specifies the person indicated by the contents of the inquiry in detail, the server (16) determines a response to the person according to the specification result, the second response determining unit (16d) And, when the second response judging unit (16d) decides to execute the dialogue with the person based on the specified result, the scenario database (26) is transmitted from the scenario database (26) via the third communication unit (16a). And a second dialogue engine unit (16e) for selecting a scenario corresponding to the dialogue;
The camera (12) includes a second speaker (12j) outputting the scenario selected by the second dialogue engine unit (16e) to the outside as a voice, and a third microphone (12b) detecting the voice of the person. A security system (10). The security system (10) according to claim 18, wherein
It is connected to the server (16), and after the end of the dialog with the person, at least the image of the person captured by the imaging unit (12a) during the dialog and the image detected by the third microphone (12b) It further has a dialogue history database (28) in which voice data of a person is recorded as dialogue history,
The second dialogue engine unit (16e) creates a new scenario using the dialogue history recorded in the dialogue history database (28) after the dialogue ends, and creates the new scenario prepared as the scenario Security system (10) characterized in that it is registered in a database (26).

Images