JP2021077245A - Number of persons for each attribute totalizing system and number of persons for each attribute totalizing device - Google Patents

Abstract

To totalize beneficial and correct information about visitors to a commercial facility and customers to tenants for each attribute in a compound commercial facility having the plurality of tenants and provide the totalized information.SOLUTION: An image analysis device comprises: a person identification unit which processes a photographed image from a camera installed in a commercial facility and identifies a photographed person from a previously registered feature amount; an attribute estimation unit which estimates the attribute of the person photographed by the camera; and a shop entry determination unit which determines that the person photographed by the camera has entered a shop. The device totalizes the number of persons who have entered a shop on the basis of the result of person identification output from the person identification unit, the result of attribute determination output from the attribute estimation unit and the result of shop entry determination output from the shop entry determination unit and provides the totalized number of persons to managers of the tenants and the commercial facility.SELECTED DRAWING: Figure 1

Description

The present invention relates to an attribute-based number counting system and an attribute-based number counting device.

Conventionally, an image captured by a camera is divided into appropriate zones, data is obtained by counting the number of people staying in the zone and the number of people entering and exiting the zone by an image processing means, and this data is associated with a POS (Point of Sale) system. A system has been proposed (see Patent Document 1).

Japanese Unexamined Patent Publication No. 2006-285409

The technique disclosed in Patent Document 1 counts the time spent in each zone and the number of people by dividing the image captured by the camera into appropriate zones on the system. However, with the technology described in Patent Document 1, the staff working in each tenant may be erroneously counted as a store entrant, and since the information about the store occupant is only the analyzed gender and age, each tenant There was a problem that it was difficult to make use of it in the operation of.

An object of the present invention is to take time and effort to manage useful and accurate information when taking measures for improving sales and attracting customers in a complex commercial facility such as a shopping mall or a department store where a plurality of tenants enter. It is an object of the present invention to provide an attribute-based number counting system and an attribute-based number counting device that can be obtained without.

In order to solve the above problems, for example, the configuration described in the claims is adopted.
The present application includes a plurality of means for solving the above problems, and to give an example thereof, the present invention connects an image analysis device into which an image captured by a camera is input, and the image analysis device by a wide area network. It is an attribute-based number counting system consisting of a center device to be used.
The image analysis device in the attribute-based number counting system of the present invention processes the image captured by the camera and identifies the captured person from the pre-registered feature amount, and the attribute of the person captured by the camera. It is provided with an attribute estimation unit for estimating the above, and a store entry determination unit for determining that a person imaged by a camera has entered the store.

Further, the center device in the attribute-based number counting system of the present invention includes a person identification result output from the person identification unit of the image analysis device, an attribute determination result output from the attribute estimation unit, and a store entry determination unit. It includes a totaling unit that counts the number of people who have entered the store based on the output result of the store entry determination, and a totaling result output unit that outputs the totaling result by the totaling unit.

According to the present invention, in a complex commercial facility such as a shopping mall or a department store where a plurality of tenants enter, it is beneficial and accurate for the tenants who open stores and the managers of the commercial facilities to improve sales and attract customers. Information can be provided.
Issues, configurations and effects other than those described above will be clarified by the description of the following embodiments.

It is a block diagram which shows the structure and function of the person number counting system by attribute in 1st Embodiment of this invention. It is a figure which shows the hardware configuration of the image processing apparatus which is the main configuration of the number of people counting system by attribute in the 1st Embodiment of this invention. It is a figure which shows the hardware configuration of the center apparatus which is the main configuration of the number of people counting system by attribute in 1st Embodiment of this invention. It is a table which shows the data structure of the attribute / person analysis result in the attribute-based number counting system of 1st Embodiment of this invention. It is a table which shows the data structure of the flow line history information in the attribute-based number counting system of 1st Embodiment of this invention. It is a figure which shows the setting example of the virtual line and the virtual area near the entrance / exit of a tenant in the attribute-based number counting system of 1st Embodiment of this invention. It is a figure which shows the setting example of the virtual line near the entrance / exit of a commercial facility in the attribute-based number counting system of 1st Embodiment of this invention. It is a flowchart which shows the flow of the attribute / person analysis processing and the flow line history creation processing in the attribute-based number counting system of 1st Embodiment of this invention. It is a block diagram which shows the structure and function of the person number counting system by attribute in the 2nd Embodiment of this invention.

<Explanation of Example of First Embodiment of the Present Invention>
Hereinafter, an example of the first embodiment for carrying out the present invention (hereinafter, referred to as “this example”) will be described with reference to FIGS. 1 to 8.
[Explanation of overall configuration]
FIG. 1 is a functional block diagram showing an attribute-based number counting system of this example. The attribute-based number counting system of this example includes the image analysis device 20 and the camera 30 installed in the commercial facility 10, and the data counting server 51, the Web server 52, and the database (DB) 53 installed in the center device 50. I have. Although one camera 30 may be installed, it is desirable that a plurality of cameras 30 are installed as shown in FIG.

The image analysis device 20 installed in the commercial facility 10 is connected to the center device 50 via the wide area network 40.
The camera 30 is attached to the commercial facility 10 toward an area where the flow line of a visitor is to be confirmed, such as the entrance / exit of a tenant opening a store in the commercial facility 10 or an elevator hall.
The image analysis device 20 includes a control unit 21, a camera management unit 22, an image analysis unit 23, a flow line analysis unit 24, a storage unit 25, and a communication unit 28.

The camera management unit 22, the image analysis unit 23, the flow line analysis unit 24, the storage unit 25, and the communication unit 28 are connected to the control unit 21. The control unit 21 controls the operation of the image analysis device 20 by executing the control program stored in the storage unit 25.
A control unit 21 and a plurality of cameras 30 are connected to the camera management unit 22. The image captured by the camera 30 is input to the camera management unit 22, and the camera management unit 22 manages the captured image for each camera 30 that has captured the captured image.

The image analysis unit 23 is connected to the control unit 21. The image analysis unit 23 includes a person detection unit 231 that detects a person in the input captured image, a person identification unit 232 that identifies the person, and an attribute estimation unit 233 that analyzes the attributes of the person.

The person detection unit 231 performs a process (person detection process) for detecting a "person having a shape presumed to be a person" existing in the captured image. The person identification unit 232 compares the person detected by the person detection unit 231 with the facial features of a plurality of pre-registered persons, and determines whether or not there is a matching or extremely similar person. Perform processing (person identification processing). The attribute estimation unit 233 performs a process (attribute analysis process) of estimating the gender and age of the person detected by the person detection unit 231 based on the face, profile, and the like.

The flow line analysis unit 24 is connected to the control unit 21. The flow line analysis unit 24 has a store entry determination unit 241 that determines that a person imaged by the camera 30 has entered the tenant store.
The store entry determination unit 241 performs a virtual line passage detection process for detecting that the target person has passed the virtual line set in advance by the tenant, that is, has entered the store, based on the continuously captured images input. Do.

In addition, the flow line analysis unit 24 performs a virtual area retention determination process indicating that the target person has stayed in the virtual area set in advance by the tenant for a predetermined period, that is, the target person is likely to enter the store. .. The details of the virtual line passage detection process will be described later in FIG. 6, but it is a process for detecting that a target person has crossed a virtual line (virtual line) preset by the tenant on the captured image.

The virtual area retention determination process is a process for detecting that a person has been staying in a virtual area preset by the tenant for a certain period of time on the captured image. The flow line analysis unit 24 performs a staff determination process that considers a person who repeatedly enters the store only for the same tenant as a tenant staff from the virtual line passage detection process and the virtual area retention determination process, and visits the store and the tenant staff. I try to distinguish between.

The storage unit 25 stores the attribute / person analysis result 26 and the flow line history information 27.
The communication unit 28 is connected to the control unit 21 and the wide area network 40. The communication unit 28 transmits the information stored in the storage unit 25 to an external device such as the center device 50 via the wide area network 40.

The center device 50 includes a data aggregation server 51, a Web server 52, and a database (DB: Data Base) 53. The data aggregation server 51 and the Web server 52 are connected to the image analysis device 20 via the wide area network 40. The database (DB) 53 is connected to the data aggregation server 51 and the Web server 52.

In the attribute-based number counting system of this example, the data counting server 51 and the Web server 52 provided in the center device 50 have a server configuration divided according to the purpose, but the server configuration is not limited to this, and a single server is configured. , A single server may have the roles of both the data aggregation server 51 and the Web server 52.

The data aggregation server 51 includes an aggregation unit 511 and an aggregation result output unit 512. The aggregation unit 511 aggregates the visitor information to the commercial facility 10 and the store visit information to each tenant obtained by the image analysis device 20, and stores this in the database (DB) 53.
At this time, the aggregation unit 511 of the data aggregation server 51 stores the information associated with the visitors using the attribute and the person ID of the person analysis information described later in FIG. 4 as keys in the database (DB) 53. The information stored in the database (DB) 53 is the number of visitors to the commercial facility 10 and each tenant, which is aggregated by location, time zone, gender and age of the person.

The aggregation result output unit 512 of the data aggregation server 51 provides the aggregation result of the aggregation unit 511 to the manager and the tenant of the commercial facility 10 via the Web server 52. As will be described later in FIG. 6, the Web server 52 is also used when each tenant sets a virtual line for determining a visitor who comes to his / her store and a virtual area in front of his / her own store.

[Hardware configuration of image analysis device 20 and center device 50]
FIG. 2 shows the hardware configuration of the image analysis device 20 used in the attribute-based number counting system of this example.
As shown in FIG. 2, the image analysis device 20 includes a CPU (Central Processing Unit) 32, a ROM (Read Only Memory) 33, a RAM (Random Access Memory) 34, and a non-volatile storage 35 connected to the bus 31. It also includes a communication interface (communication IF) 36 and an input unit 37 for communicating with the outside.

A plurality of cameras 30 of the commercial facility 10 shown in FIG. 1 are connected to the input unit 37 of FIG. 2, and the image captured by the cameras 30 is stored in the non-volatile storage 35 via the bus 31.
The CPU 32 reads the program code of the software that realizes the functions of the camera management unit 22, the image analysis unit 23, the flow line analysis unit 24, and the like of the image analysis device 20 (FIG. 1) from the ROM 33 and executes the program code.

Variables and the like generated during the arithmetic processing performed in the image analysis apparatus 20 are temporarily written in the RAM 34. The CPU 32 realizes the function of the control unit 21 of FIG. 1 by executing the program code recorded in the ROM 33.

The communication interface (I / F) 36 corresponds to the communication unit 28 shown in FIG. 1, and as the communication I / F 36, for example, a NIC (Network Interface Card) or the like is used. Through this communication I / F 36, the image analysis device 20 communicates with the data aggregation server 51 and the Web server 52 via the wide area network 40 shown in FIG. The wide area network 40 is a wide area communication network represented by the Internet, LTE (Long Term Evolution), LAN (Local Area Network), etc., and the image analysis device 20 is a data aggregation server via the wide area network 40. Various data are exchanged with the 51 and the Web server 52.

The non-volatile storage 35 is composed of a non-volatile memory such as an SSD (Solid State Drive), and stores and stores programs, data, and the like necessary for operating the CPU 32. The non-volatile storage 35 is provided with storage areas for the attribute / person analysis result 26 and the flow line history information 27 in the storage unit 25 shown in FIG.

FIG. 3 shows the hardware configuration of the center device 50 of the attribute-based number counting system of this example. Specifically, the hardware configuration in which the data aggregation server 51 and the database (DB) 53 are connected, and the hardware configuration in which the Web server 52 and the database (DB) 53 are connected are shown.
The center device 50 also includes a CPU 42, a ROM 43, a RAM 44, a non-volatile storage 45 (corresponding to the database (DB) 53 in FIG. 1), and a communication unit I / F 46 connected by a bus 41. Since these configurations and functions are the same as the hardware configurations and functions of the image analysis apparatus 20 described with reference to FIG. 2, the description thereof will be omitted.
However, the center device 50 does not have a configuration corresponding to an input unit 37 that receives an image pickup signal from the camera 30 shown in FIG.

[Description of data structure table]
FIG. 4 shows a table of the data structure of the attribute / person analysis result 26 in the present embodiment, which is stored in the storage unit 25 of the attribute-based number counting system of this example shown in FIG. The attribute / person analysis result 26 has an item of person ID 26a, an item of name 26b, an item of gender 26c, an item of age 26d, an item of estimation accuracy 26e, and an item of person identification situation 26f.

In the item of the person ID 26a, ID information of a unique character string that uniquely identifies the record of the attribute / person analysis result is stored. This ID information is information assigned to each person detected by the person detection process performed by the image analysis unit 23.
In the item of the name 26b, the pre-registered name is stored when it is determined by the person identification process performed by the image analysis unit 23 that the captured person matches or is extremely similar to the pre-registered information.
In the item of gender 26c, the gender estimated from the photographed person by the attribute analysis process performed by the image analysis unit 23 is stored.

In the item of age 26d, the age estimated from the photographed person by the attribute analysis process performed by the image analysis unit 23 is stored.
The item of estimation accuracy 26e was estimated from the captured image which was the basis of estimation such as the angle and resolution of the captured image when the gender and age were estimated from the captured person by the attribute analysis process performed by the image analysis unit 23. The accuracy of the result is stored.
In the item of the person identification situation 26f, information indicating whether or not the photographed person is determined to match or be extremely similar to the information registered in advance by the person identification process performed by the image analysis unit 23 is stored.

FIG. 5 is a table showing the data structure of the flow line history information 27, which is also stored in the storage unit 25 in the attribute-based number counting system of this example shown in FIG. The flow line history information 27 has an item of the person ID 27a, an item of the date and time 27b, and an item of the place 27c.
In the item of the person ID 27a, ID information assigned to each person detected by the person detection process performed by the image analysis unit 23 is stored.

In the item of the date and time 27b, the date and time when the person ID 27a enters the tenant or passes through the specific area by the virtual line passage detection process or the virtual area retention determination process performed by the flow line analysis unit 24 is stored.
In the item of the place 27c, the name of the tenant in which the person ID 27a has entered or the name of the specific area in which the person ID 27a has passed is stored by the virtual line passage detection process or the virtual area retention determination process performed by the flow line analysis unit 24.

[Example of setting virtual line / virtual area]
FIG. 6 shows an example of setting a virtual line / virtual area near the entrance of a tenant in the attribute-based number counting system of this example. The virtual line and the virtual area are set by the administrator or the tenant of the commercial facility 10 on the Web screen provided by the Web server 52 of the center device 50. The information of the virtual line and the virtual area set by the Web server 52 is transmitted from the center device 50 to the image analysis device 20.

In addition, when setting the virtual line and the virtual area on the Web screen of the Web server 52, the manager or the tenant of the commercial facility 10 refers to the image captured by the camera 30 installed in the commercial facility 10 on the image capture screen. Set.

FIG. 6 is an example in which tenant A to tenant E set their own virtual lines and virtual areas on the captured image of the camera 30 overlooking the entrance / exit of tenant E from tenant A. In this example, tenant A, tenant C, and tenant E set only virtual lines as the entrance / exit of the tenant. Tenant B and tenant D set a virtual line at the entrance / exit of the tenant and a virtual area at the storefront.

When the person a crosses the virtual line set by the tenant C in the direction of entering the store, the person a is considered to have entered the tenant C. Further, if the person b continues to stay in the virtual area set by the tenant D for a certain period of time, the person b is considered to have entered the tenant D.

FIG. 7 shows an example in which a virtual line is set near the entrance / exit of the commercial facility 10 in the attribute-based number counting system of this example. That is, this is an example in which the manager of the commercial facility 10 sets a virtual line on the captured image of the camera 30 overlooking the entrance / exit of the commercial facility 10 in order to grasp the total number of visitors. In this example, virtual lines are set for each of the two entrances / exits 1 and 2 of the commercial facility 10. When the person c crosses the virtual line set at the entrance 1 of the commercial facility 10 into the commercial facility 10, the person c is considered to have visited the commercial facility 10.

As described above, as described with reference to FIGS. 6 and 7, the number of visitors to each tenant and the number of visitors to the commercial facility 10 can be grasped by using the virtual line and the virtual area to enter the commercial facility 10. It is possible to grasp what percentage of the total number of visitors who visited each tenant entered the store.

[Processing procedure of the number of people counting system by attribute in this example]
Next, the operation of the attribute-based number counting system according to this embodiment will be described with reference to the drawings. FIG. 8 is a flowchart showing an image analysis process performed by the image analysis apparatus 20 in this example.
The process performed by the attribute-based number counting system of this example is a process of creating and updating records related to the attribute / person analysis result 26 shown in FIG. 4 and the flow line history information 27 shown in FIG. This process is performed based on the captured image captured by the camera 30.

When the image analysis device 20 receives the captured image from the camera 30, the process shown in FIG. 8 is started.
When the process is started, the person detection unit 231 of the image analysis unit 23 performs the person detection process based on the received captured image (step S101). The processing from step S102 to step S110 is a repetitive processing (loop processing). This iterative process is repeated for the number of persons detected in step S101.

First, person information is acquired one by one from the list of people detected in the captured camera image (step S102). In this step S102, it is determined whether or not the nth person information in the person list exists in the attribute / person analysis result (see FIG. 4) of the storage unit 25. That is, when the nth person information in the person list detected in the camera image is present in the attribute / person analysis result shown in FIG. 4, the person detection unit 231 of the image analysis unit 23 is shown in FIG. In the person identification situation 26f shown, it is determined whether the nth person is confirmed or undetermined (step S103). Here, the camera image may be an image captured by one camera 30, an image captured by a plurality of cameras 30, or an image obtained by synthesizing these images.

That is, in this step S103, the attribute / person analysis result of FIG. 4 is searched based on the person ID of the nth person information, and the content stored in the item of the person identification status 26f of the uniquely specified record. Will have been acquired. When the nth person information does not exist in the person list of the camera image detected in step S101, that is, for all the people (n-1) in the camera image, the processes of steps S102 to S110 are completed. If so, the iterative process of steps S102 to S110 ends. Note that n is a natural number starting from 1.

If the content stored in the item of the person identification status 26f in step S103 is "undecided" (Yes in step S103), the attribute estimation unit 233 of the image analysis unit 23 performs the attribute analysis process of the nth person information. (Step S104). That is, in the attribute analysis process of step S104, gender estimation and age estimation for the nth person information are performed, and the estimation accuracy is calculated. If the gender and age cannot be estimated, it is considered as "unknown".

Then, the attribute estimation unit 233 stores the obtained gender, age, and estimation accuracy of the nth person in the attribute / person analysis result 26 of the storage unit 25 of FIG. This stored result is as shown in FIG. However, in step S104, when information other than "unknown" is stored as the estimation result and the estimated accuracy already stored is larger than the estimated accuracy calculated this time, the information of the estimated accuracy calculated this time is used. To be stored with priority.

Following the attribute analysis process in step S104, the person identification unit 232 of the image analysis unit 23 performs a person identification process by face recognition on the nth person information in the camera image (step S105). In the person identification process of step S105, the person identification unit 232 registers the registered information of the member information image of the visitor and the image of the store staff information registered in advance with respect to the nth person information in the camera image. Based on this, the match or mismatch with the registered information is estimated.

Then, in step S105, when the nth person information in the camera image matches the registered information and the person can be identified, the name, gender, and age are stored from the registered information in FIG. 1. It is stored in the attribute / person analysis result 26 of the part 25. At that time, based on the person ID of the nth person information in the camera image, the database of the attribute / person analysis result in FIG. 4 is searched, and "confirmed" is entered in the item of the person identification status 26f in FIG. Store.

If it is determined in step S103 that the content stored in the item of the person identification status 26f is not "unconfirmed", that is, "confirmed" (No in step S103), then the attribute estimation unit 233 determines whether or not the nth person in the camera image is a general visitor (step S111). If it is determined in step S111 that the store is not a general visitor, for example, if it is determined that the store is a staff member of a tenant store (No in step S111), the process proceeds to step S110 and the process ends.

If the person is identified from the nth person information in the camera image in step S105, or if it is determined to be a general visitor in step S111, the flow line analysis unit 24 subsequently determines. , The flow line analysis process is performed on the nth person information in the camera image (step S106).
In the flow line analysis process of step S106, a virtual line passage detection process, a virtual area retention determination process, and a staff determination process are performed.

First, for the nth person information in the camera image, whether or not the passage of the virtual line set in advance by each tenant is detected, or whether the nth person stays in the virtual area for a certain period of time. It is determined whether or not it has been detected (step S107).

Then, in step S107, it is determined that the passage of the virtual line is detected or the residence for a certain period of time is detected in the virtual area for the nth person information in the camera image (Yes in step S107). Next, the flow line analysis unit 24 determines whether or not the person who has passed the virtual line or the person who has been detected to stay in the virtual area for a certain period of time is not a staff member (step S108).

If it is determined in step S108 that the nth person who has passed through the virtual line or the nth person who has been detected to stay in the virtual area for a certain period of time is not a staff member (Yes in step S108), the flow line The analysis unit 24 performs a flow line history registration process on the nth person information in the camera image (step S109). In the flow line history registration process of step S109, additional records are registered in the flow line history information 27 of the storage unit 25 of FIG. 1 based on the virtual line passage detection or the virtual area retention determination detected in step S106 ( Step S110).

This additional registration corresponds to the additional registration of a record for the flow line history information of FIG. That is, the flow line analysis unit 24 performed the person ID of the nth person information in the camera image, the date and time when the virtual line passage detection or the virtual area retention determination was performed, and the virtual line passage detection or the virtual area retention determination. The location is stored in the flow line history information 27 (see FIG. 5) of the storage unit 25 of FIG.

When it is determined in step S107 that the passage through the virtual line is not detected or the residence for a certain period of time is not detected in the nth person information in the camera image (No in step S107). , And in step S108, when it is determined that the person who has passed the virtual line or the nth person who has been detected to stay in the virtual area for a certain period of time is a staff member (No in step S108), steps S103 to The loop processing of S109 is terminated.

As described above, according to the image analysis processing flow shown in FIG. 8, the analysis result of the captured image is stored in the attribute / person analysis result 26 shown in FIG. 4 and the flow line history information 27 shown in FIG.
The data aggregation server 51 of the center device 50 periodically acquires the information stored in the attribute / person analysis result 26 shown in FIG. 4 and the flow line history information 27 shown in FIG. It is stored in 60 databases (DB) 53.

In the aggregation process by the data aggregation server 51, for example, information obtained by combining the flow line history information of FIG. 5 with the person ID of the attribute / person analysis result of FIG. 4 as a key is stored in the database (DB) 53. The information stored in the database (DB) 53 is counted for each location and each time zone, and the number of visitors for each time zone of each tenant can be obtained. Furthermore, the number of visitors by attribute can be obtained by dividing the visitors by gender and age. In addition, each person ID who visits the own tenant counts the other tenants who have visited by tenant, so that the flow line tendency of the visitors who come to the own tenant can be obtained.

<Explanation of Example of Second Embodiment of the Present Invention>
FIG. 9 is a block diagram showing a configuration and a function of an attribute-based number counting system according to a second embodiment of the present invention.
The difference from the first embodiment shown in FIG. 1 is that in the first embodiment, the data aggregation server 51 of the center apparatus 50 has the aggregation unit 511 and collects various data transmitted from the image analysis apparatus 20. On the other hand, in the second embodiment shown in FIG. 9, the aggregation unit 29 is provided in the image analysis device 20.

Therefore, the data server 51'of the center device 50 in the second embodiment includes only the aggregation result output unit 512 of the data aggregation server 51 of the first embodiment. Since all other configurations are the same, duplicate explanations will be omitted here.

It can be considered that the aggregation unit 29 of FIG. 9 has transferred the function of the aggregation unit 511 of the data aggregation server 51 of the embodiment of FIG. 1 to the image analysis device 20 side. Therefore, the aggregation unit 29 in FIG. 9 counts the information associated with the visitors by using the attribute and the person ID of the person analysis information described later in FIG. 4 as keys for each place, each time zone, and each person's gender and age. And store it in the storage unit 25. The analysis information regarding the aggregation result is provided to the manager and the tenant of the commercial facility 10 by the Web screen of the Web server 52.

According to the first and second embodiments of the present invention, the tenant who opens a store in the commercial facility 10 can be provided with the number of visitors for each attribute for each time zone. In addition, it is possible to receive information on the flow line tendency of visitors, such as which direction people tend to enter the store, and many visitors enter the store after visiting a specific store.

In addition, the manager of the commercial facility 10 can grasp the transition of the number of visitors to each tenant, and thereby it is possible to estimate the exit of the tenant. Furthermore, since system settings such as virtual line / virtual area settings and store staff registration can be entrusted to each tenant, there is an advantage that management effort is small.

The present invention is not limited to the above-described embodiment, and includes various modifications. The above-described embodiments have been described for the purpose of explaining the present invention in an easy-to-understand manner, and the present invention is not necessarily limited to those having all the described configurations, and the present invention is not necessarily limited to those having all the described configurations. It can be applied.

Further, each of the above configurations, functions, processing units, processing means and the like may be realized by hardware by designing a part or all of them by, for example, an integrated circuit. Further, each of the above configurations, functions, and the like may be realized by software by the processor interpreting and executing a program that realizes each function. Information such as programs, tables, and files that realize each function can be placed in a memory, a hard disk, a recording device such as an SSD (Solid State Drive), or a recording medium such as an IC card, an SD card, or a DVD.

Further, in the drawings, the control lines and information lines are shown as necessary for explanation, and not all the control lines and information lines are necessarily shown in the product. In practice, almost all configurations may be considered interconnected.

10 ... commercial facility, 20 ... image analysis device, 21 ... control unit, 22 ... camera management unit, 23 ... image analysis unit, 231 ... person detection unit, 232 ... person identification unit, 233 ... attribute estimation unit, 24 ... movement line Analysis unit, 241 ... Store entry judgment unit, 25 ... Storage unit, 26 ... Attribute / person analysis result, 27 ... Movement line history information, 28 ... Communication unit, 30 ... Camera, 40 ... Wide area network, 50 ... Center device, 51 … Data aggregation server, 511… Aggregation department

Claims (6)

It is an attribute-based number counting system composed of an image analysis device for inputting an image captured by a camera and a center device connected to the image analysis device by a wide area network.
The image analysis device is
A person identification unit that processes an image captured by the camera and identifies the captured person from a pre-registered feature amount, and a person identification unit.
An attribute estimation unit that estimates the attributes of a person captured by the camera,
It is equipped with a store entry determination unit that determines that a person photographed by the camera has entered the store.
The center device is
Based on the person identification result output from the person identification unit of the image analysis device, the attribute determination result output from the attribute estimation unit, and the store entry determination result output from the store entry determination unit. A tabulation department that counts the number of people who have entered the store,
A total result output unit that outputs a total result by the total unit is provided.
Number of people counting system by attribute. The store entry determination unit refers to one or a plurality of stores imaged by one or a plurality of cameras.
The attribute-based number counting system according to claim 1, wherein it is determined that a person photographed by the one or more cameras has entered the store. The image analysis device further
A virtual virtual line indicating whether or not a store is entered is set for each tenant, and based on the continuously captured images input, it is detected that a person captured by the camera has crossed the virtual line. The attribute-based number counting system according to claim 2, further comprising a movement line analysis unit for determining the entry of the person by the store entry determination unit. The third aspect of claim 3, wherein the flow line analysis unit sets a virtual area retention area near the store of each tenant on the captured image, and detects that a person has been staying in the virtual area retention area for a certain period of time. Number of people counting system by attribute. The attribute estimation unit of the image analysis device determines whether or not the imaged person is a store staff, and determines whether or not the person is a store staff.
The totaling unit of the center device totals the number of visitors excluding the store staff determined by the attribute estimation unit for each attribute and for each store.
The attribute-based number counting system according to claim 1, wherein the totaling result output unit outputs the totaled results by the totaling unit. A person identification unit that processes the image captured by the camera and identifies the captured person from the pre-registered features.
An attribute estimation unit that estimates the attributes of a person captured by the camera,
A store entry determination unit that determines that a person photographed by the camera has entered the store,
The number of people who entered the store based on the person identification result output from the person identification unit, the attribute determination result output from the attribute estimation unit, and the store entry determination result output from the store entry determination unit. It is equipped with a counting unit that counts.
Number of people counting device by attribute.

Images