JP2009193424A - Traffic line capturing device and traffic line capturing method and program - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To precisely capture the traffic line of a user without making a user carry any special equipment such as an active RFID tag. <P>SOLUTION: A collation means 5 collates the image of a user photographed by card reader interlocking cameras LC1 to LC3 with images photographed by a plurality of cameras C11 to C34 whose photographic positions are different, and when determining that there exists an image in which the user is projected among images photographed by the plurality of cameras C11 to C34, stores records including the photographic time, photographic position and accuracy of the image in a path information table 4. A narrowing-down means 6 deletes the records in which the user is not able to move by the photographic time included in the records to the photographic position included in the records among the records stored in the path information table 4. <P>COPYRIGHT: (C)2009,JPO&INPIT

Description

  The present invention relates to a flow line capturing technique for capturing a user's flow line, and more particularly, to a flow line capturing technique that can accurately capture a user's flow line without having the user carry a special device.

  Various techniques are known for capturing user traffic lines, such as customer traffic lines in stores and passenger traffic lines in airports. For example, non-contact recognition such as active RFID tags. There is known a method of allowing a user to carry a possible device and installing a reader for reading an active RFID tag in various places such as in a store and capturing a user's flow line based on the reading result of the reader. However, this method has a problem that a user must carry a special device such as an active RFID tag.

  A technique is also known in which a user's flow line is captured based on a user's face image (see, for example, Patent Document 1). The technique described in Patent Document 1 installs cameras at various locations in an airport, and collates images taken by them with passenger images taken in advance, thereby The flow line is captured. That is, the user's face image and each image captured by each camera are collated to detect an image of the user, and by connecting the installation position of the camera that captured the detected image, the user's The flow line is captured.

Re-specialized WO2005 / 055151

  According to the technique described in Patent Document 1, it is possible to capture a user's flow line without having the user carry a special device such as an active RFID tag. However, since the technique described in Patent Document 1 detects a user's image by collating the user's face image with each image captured by each camera, There is a problem that the line cannot be captured with high accuracy. In other words, it is difficult to accurately detect an image in which a user's face image is captured, and an image in which a face image other than the user is captured may be erroneously detected. It was difficult to do.

(Object of invention)
Accordingly, an object of the present invention is to provide a flow line capturing device that can accurately capture a user's flow line without having the user carry a special device.

The first flow line capturing device according to the present invention is:
Multiple cameras with different shooting positions,
A route information storage unit;
When the user's reference person image is collated with images taken by the plurality of cameras, and it is determined that there is an image of the user in the images taken by the plurality of cameras, Collating means for storing route information including a photographing time and a photographing position in the route information storage unit;
Of the route information stored in the route information storage unit, delete the route information that the user cannot move to the shooting position included in the route information by the shooting time included in the route information. And a narrowing-down means.

A first flow line capturing method according to the present invention includes:
The computer collates the reference person image of the user with images taken by a plurality of cameras having different shooting positions, and determines that there is an image showing the user in the images taken by the plurality of cameras. A matching step for storing path information including the shooting time and shooting position of the image in the path information storage unit;
The computer cannot move to the shooting position included in the route information among the route information stored in the route information storage unit by the shooting time included in the route information. And a narrowing step for deleting route information.

The first program according to the present invention is:
A program for causing a computer including a plurality of cameras having different shooting positions and a path information storage unit to function as a flow line capturing device,
The computer,
When the user's reference person image is collated with images taken by the plurality of cameras, and it is determined that there is an image of the user in the images taken by the plurality of cameras, Collating means for storing route information including a photographing time and a photographing position in the route information storage unit;
Of the route information stored in the route information storage unit, delete the route information that the user cannot move to the shooting position included in the route information by the shooting time included in the route information. Function as a narrowing-down means.

  According to the present invention, it is possible to accurately capture a user's flow line without having the user carry a special device.

  Next, the best mode for carrying out the present invention will be described in detail with reference to the drawings.

Embodiment of the present invention
Referring to FIG. 1, a flow line capturing device according to an embodiment of the present invention includes a customer information input means 1, a customer information database (DB) 2, a customer specifying means 3, a route information table 4, and a matching means. 5, narrowing means 6, floor map DB 7, route determination means 8, multiple card readers R 1 to R 3, multiple card reader linked cameras LC 1 to LC 3, and multiple cameras C 11 to C 34. I have.

  The customer information input means 1 is a means for registering personal information such as the customer's name and age and the customer ID in the customer information DB 2, and usually a keyboard for manually inputting data, handwritten information This is realized by OCR or the like for inputting.

  The card readers R1 to R3 have a function of reading a customer ID such as a membership number recorded on a card carried by the customer, a function of sending the read customer ID and the device number of the reader to the customer specifying means 3, It has a function of sending the read customer ID to the card reader linked cameras LC1 to LC3.

  2, 4, and 5 are floor views of the first floor, the second floor, and the third floor of the store, respectively. The card reader R1 is installed near the entrance of the first floor, and the card readers R2 and R3 are installed on the second and third floors, respectively. The card carried by the customer may be a card printed with a customer ID, coded and printed, recorded as magnetic data, or recorded on a non-contact IC such as an RFID tag or FeliCa. it can. The card readers R1 to R3 correspond to the recording form of the customer ID, and those printed with characters are read by the optical reader, and then the character ID is recognized to recognize the customer ID. For barcodes, QR codes, etc., the code is read by an optical reader. In the case of a magnetic card, read with a magnetic card reader. In the case of a non-contact IC, it is read by an RFID reader or a non-contact IC card reader.

  Customers who come to the store can enjoy various services provided by the store by having the card readers R1 to R3 read the card carried by the card reader. For example, by making the card reader R1 read the card, it is possible to acquire store visit points or to acquire information in the store (for example, by reading the card, an information mail is delivered to the mobile phone from the store). Further, by causing the card readers R2 and R3 to read the card, it is possible to acquire product information for each sales floor and to make reservations for tickets and restaurants. What kind of service the store side performs can be set for each card reader.

  When the customer ID and the device number are sent from the card readers R1 to R3, the customer specifying means 3 generates a route information table 4 corresponding to the customer ID on a storage device such as a disk device, and at the head thereof , The shooting position of the card reader-linked camera corresponding to the sender card reader (which can be recognized from the device number) as the shooting position, the current time as the shooting time, and a record containing 100% as the degree of probability of being the person Record. For example, if a customer ID is sent from the card reader R1 at time T1, a record including the shooting position “P1”, the shooting time “T1”, and the accuracy “100%” shown in FIG. 2 is recorded. If the route information table 4 corresponding to the customer ID sent from the card reader has already been generated, a record composed of the shooting position, shooting time, and accuracy is added to the already generated route information table 4. Register additional. In addition, the shooting position of each card reader-linked camera LC1 to LC3 can be set to the center position of the floor that can be shot by each of the card reader-linked cameras LC1 to LC3, for example.

  As shown in FIG. 3, each of the card reader-linked cameras LC1 to LC3 is installed so that the card reader R1 to R3 can take a picture of the customer, and the customer is sent from the card reader R1 to R3. Using the ID as a trigger, it has a function of taking an image of the moment when the customer reads the card with the card readers R1 to R3, and sending the taken image and the customer ID as a pair to the collating means 5.

  The cameras C <b> 11 to C <b> 34 have a function of capturing a predetermined number of frames per unit time and transmitting the captured image, the capturing time, and its own device number to the verification unit 5.

  The collation means 5 collates the images captured by the card reader-linked cameras LC1 to LC3 with the images captured by the cameras C11 to C34, and an image in which the customer is copied in the images captured by the cameras C11 to C34. Is present, the path information (record) including the shooting time, shooting position, and accuracy of the image is stored in the path information table 4. The collation means 5 having such a function includes a reference feature extraction means 51, a reference feature DB 52, a person feature extraction means 53, a person feature DB 54, and an accuracy detection means 55.

  When a pair of image and customer ID is sent from the card reader interlocking cameras LC1 to LC3, the reference feature extraction means 51 extracts a person image from the image, and further extracts feature information from the extracted person image. The feature information is stored in the reference feature DB 52 in association with the customer ID. As the feature information, for example, facial feature points, clothing color patterns, and the like can be used.

  When the images, device numbers, and shooting times are sent from the cameras C11 to C34, the person feature extraction means 53 extracts a person image from the image, further extracts feature information from the extracted person image, and the feature information. Is stored in the person feature DB 54 in association with the device number and the photographing time.

  The accuracy detection unit 55 compares the feature information of each customer stored in the reference feature DB 52 with the feature information of each person image stored in the person feature DB 54, so that each person is It has a function of calculating the probability that the image is a person image of the customer, and determining that a person image having an accuracy equal to or higher than a threshold (for example, 20%) is the person image of the customer. Various methods can be used as the accuracy calculation method. For example, each item of feature information (for example, when a feature point of a face is used as feature information, the position of eyes in the face, the left and right eyes) A method of calculating the degree of coincidence of each item) and using the integrated average of the calculated degrees of coincidence of each item, or a method of obtaining the accuracy by multiplying the calculated degree of coincidence of each item be able to. Also, when a clothing color pattern is used as the feature information, the accuracy can be obtained by the same method.

  Further, the accuracy detection means 55 generates a record including the shooting position, shooting time, and accuracy of an image including a person image with accuracy equal to or greater than a threshold for each customer, and the generated record is used as the route information of the customer. Store in table 4. In the present embodiment, in order to reduce the number of records stored in the route information table 4 and to effectively use storage resources, the generated records are grouped into groups according to the customer's stay location. Only the record with the highest accuracy is stored in the route information table 4. For example, when the customer enters the shooting range of the camera C12 at the time T1 and moves from the shooting range of the camera C12 to the shooting range of the camera C13 at the time T2, with respect to the times T1 to T2, a record related to the image shot by the camera C12. Among them, the record with the highest accuracy is stored in the route information table 4.

  A floor layout diagram is registered in the floor map DB7.

  The narrowing-down means 6 uses the floor layout diagram stored in the floor map DB 7 and is included in the record to the shooting position included in the record among the records stored in the route information table 4. It has a function of deleting a record that the customer cannot move by the shooting time.

  The route determination unit 8 has a function of outputting customer movement flow line information 9 in which shooting positions stored in the route information table 4 that has been narrowed down by the narrowing-down unit 6 are arranged in time series.

  The customer specifying means 3, the collating means 5, the narrowing means 6 and the route determining means 8 can be realized by a computer, and when realized by a computer, for example, as follows. A disk, a semiconductor memory, and other recording media on which a program for causing the computer to function as the customer specifying means 3, the collating means 5, the narrowing means 6, and the route determining means 8 are prepared, and the computer reads the above programs. The computer controls its own operation according to the read program, thereby realizing the customer specifying means 3, the collating means 5, the narrowing means 6 and the route determining means 8 on the computer.

[Description of Operation of Embodiment]
Next, the operation of the present embodiment will be described in detail.

  Now, for example, it is assumed that a customer A visits a store and has his / her card read by the card reader R1 on the first floor. When the card reader R1 reads the customer ID “#A” from the card of the customer A, the card reader R1 transmits the read customer ID “#A” and the device number “No. A” of the own reader R1 to the customer identification unit 3.

  Thereby, the customer specifying means 3 searches a storage device (not shown) and checks whether or not the route information table 4 for the customer A exists. If not, the route information table 4 for the customer A is generated on the storage device, and at the beginning, the shooting position P1 of the card reader linked camera LC1 linked to the card reader R1 as the shooting position, the shooting time Record including the current time and “100%” as accuracy. On the other hand, when it exists, the record of the said content is recorded on the route information table 4 for the customer A already produced | generated. In this example, since the route information table 4 for the customer A does not exist, the customer specifying means 3 generates the route information table 4 for the customer A and then stores the record with the record number “1” in FIG. Will do.

  On the other hand, when the card reader R1 reads the card of the customer A, the card reader interlocking camera LC1 takes an image of the customer A as shown in FIG. Accordingly, the reference feature extraction unit 51 extracts a person image from the image, further obtains feature information of the extracted person image, associates the feature information with the customer ID “#A” of the customer A, and the reference feature DB 52. Register with.

  Thereafter, when the customer A moves in the store, the customer A is photographed by any of the cameras C11 to C34 installed in various places in the store. When the images, device numbers, and shooting times are sent from the cameras C11 to C34, the person feature extraction unit 53 extracts a person image from the image, further extracts feature information from the extracted person image, and the feature information. Are stored in the person feature DB 54 in association with the device number and the photographing time. In addition, the accuracy detection unit 55 compares the feature information of the customer A stored in the reference feature DB 52 with the feature information of each person image stored in the person feature DB 54 as necessary. The probability that each person image is the customer A is calculated, and the person image having the accuracy equal to or higher than the threshold is determined to be the customer A person image. Thereafter, the accuracy detection means 55 generates a record including the shooting position, shooting time, and accuracy of the image including the person image with the accuracy equal to or greater than the threshold value for the customer A, and further, the generated record The group is divided into groups according to the stay position, and only the record having the highest accuracy in each group is stored in the route information table 4 for the customer A. Thereby, the records of the record numbers 2 to 6 shown in FIG. 6 are stored in the route information table 4 for the customer A.

  Thereafter, it is assumed that the customer A causes the card reader R2 installed on the second floor shown in FIG. 4 to read his / her card.

  The card reader R2 transmits the customer ID “#A” read from the customer A card and the device number of the card reader R2 to the customer specifying means 3. When the customer ID “#A” of the customer A is sent, the customer specifying means 3 searches a storage device (not shown) and checks whether the route information table 4 for the customer A exists. In the case of this example, since the route information table 4 for the customer A already exists, the customer specifying means 3 is linked with the card reader R2 as the photographing position in the already generated route information table 4 for the customer A. A record including the shooting position P7 of the card reader-linked camera LC2, the current time “13:25” as the shooting time, and “100%” as the accuracy is stored (record number 7 in FIG. 6).

  Thereafter, when the customer A moves in the store, the person feature extraction means 53 and the accuracy detection means 55 perform the same processing as described above, and the record number shown in FIG. 8-10 records are stored.

  Next, the operation of the narrowing means 6 will be described. When the customer ID “#A” of the customer A is input from the input device such as a keyboard (not shown), for example, by an administrator or the like, the narrowing-down means 6 selects the route information table 4 for the customer A (customer ID “#A” Among the records stored in the path information table (registered in the storage device in association with ”), the customer cannot move to the shooting position included in the record by the shooting time included in the record. Delete possible records. The operation of the narrowing means 6 will be described in detail with reference to the flowchart of FIG.

  The narrowing-down means 6 reads records stored in the route information table 4 for customer A in order from the first record (step S71), and reads a record whose accuracy is not “100%” (step S73 is N). A difference Δt = t2−t1 between the photographing time t2 set in the record read this time and the photographing time t1 set in the record read last time is obtained (step S74), and is set in the record read this time. The distance between the shooting position and the shooting position set in the previously read record is obtained based on the floor layout registered in the floor map DB 7 (step S75).

  After that, the narrowing-down means 6 determines whether or not the distance obtained in step S75 can be walked and moved in the time (difference Δt) obtained in step S74 (step S76). This determination is performed, for example, by comparing the time ta obtained by dividing the distance obtained in step S75 by a predetermined general walking speed with the time Δt obtained in step S74. That is, if Δt ≧ ta, it is determined that it can move by walking, and if Δt <ta, it is determined that it cannot move.

  If it is determined that movement is possible (Y in step S76), the next record is read. On the other hand, if it is determined that the movement is impossible (N in step S76), the record with the lower set accuracy is deleted from the record read this time and the record read last time (step S77). Thereafter, the next record is read (step S71). The narrowing-down means 6 repeats the above processing until there are no unprocessed records (until step S72 becomes Y). In the present embodiment, the record having the lower accuracy of the record read last time and the record read this time is deleted, but the record read this time may be deleted.

  By performing the above-described processing on the route information table 4 for the customer A shown in FIG. 6, the contents of the route information table 4 are changed to, for example, those shown in FIG. In the path information table 4 shown in FIG. 8, records with record numbers 2, 5, and 9 are deleted.

  When the processing of the narrowing-down means 6 is completed, the route determination means 8 has customer movement flow line information 9 (P1 → P2 → P4 → P5) in which the shooting positions set in the route information table 4 for the customer A are arranged in time series. → P7 → P8 → P10) is generated and output. The route determination means 8 also performs processing for displaying the flow line of the customer A on the floor diagrams of each floor, as shown in FIGS. 9 to 11, based on the customer movement flow line information 9. 9 to 11 show floor diagrams from the first floor to the third floor. By viewing this display, customer A receives the service using the card with the card reader R1, moves to the second floor using the escalator, and further uses the card with the card reader R2 on the second floor. After that, you can see that you got out of the store by moving to the first floor using the stairs. In addition, it can be seen that the customer did not stop on the third floor.

[Effect of the embodiment]
According to the present embodiment, it is possible to accurately capture a user's flow line without causing the user to carry a special device. The reason is that the user's images taken by the card reader-linked cameras LC1 to LC3 are collated with images taken by a plurality of cameras C11 to C34 having different shooting positions, and taken by the plurality of cameras C11 to C34. If it is determined that there is an image in which the user is photographed, the collation means 5 that stores a record including the photographing time and photographing position of the image in the route information table 4 and the route information storage unit 4 Because there is a narrowing means 6 that deletes records that the user cannot move to the shooting position included in the record by the shooting time included in the record. is there. That is, in the present embodiment, the user's flow line is captured in consideration of whether the user can move as well as the image comparison result, so the user's flow line can be captured with high accuracy. Can do.

  Furthermore, according to the present embodiment, there is an effect that it is not necessary to prepare a user image in advance. The reason is that the card reader-linked camera LC1 is installed in the vicinity of the card readers R1 to R3 that read the customer ID from the user's card and takes pictures of the user when the customer IDs are read by the card readers R1 to R3. This is because it has ~ LC3.

  Moreover, in this Embodiment, about two records of the records stored in the path information storage unit 4, the shooting time included in one record and the shooting time included in the other record If the difference is shorter than the time required to move the distance between the shooting position included in the one record and the shooting position included in the other record, the difference between the two records Since the route information with the lower accuracy is deleted, the user's flow line can be captured with extremely high accuracy.

  The present invention is preferably applied to capturing a flow line of a customer in a store, capturing a flow line of a passenger in an airport, capturing a flow line of a worker in a limited area, and the like.

It is a block diagram which shows the structural example of the flow-line capture apparatus concerning embodiment of this invention. It is a floor figure of the 1st floor. It is a figure which shows the positional relationship of card reader R1-R3 and card reader interlocking | linkage cameras LC1-LC3. It is a floor figure of the 2nd floor. It is a floor figure of the 3rd floor. It is a figure which shows the example of the content of the path | route information table. 4 is a flowchart showing a processing example of narrowing means 6; It is a figure which shows the example of the content of the path | route information table 4 after the process by the narrowing-down means 6 was implemented. It is a figure which shows the display result of the customer's flow line in the 1st floor. It is a figure which shows the display result of the customer's flow line in a 2nd floor. It is a figure which shows the display result of the customer's flow line in the 3rd floor.

Explanation of symbols

1 ... Customer information input means 2 ... Customer information DB (database)
3 ... customer specifying means 4 ... route information table 5 ... matching means 51 ... reference feature extracting means 52 ... reference feature DB (database)
53 ... person feature extraction means 54 ... person feature DB (database)
55 ... Accuracy detection means 6 ... Refinement means 7 ... Floor map DB (database)
8 ... Route determination means 9 ... Customer movement flow line information R1-R3 ... Card reader LC1-LC3 ... Card reader linked camera C11-C34 ... Camera

Claims (15)

Multiple cameras with different shooting positions,
A route information storage unit;
When the user's reference person image is collated with images taken by the plurality of cameras, and it is determined that there is an image of the user in the images taken by the plurality of cameras, Collating means for storing route information including a photographing time and a photographing position in the route information storage unit;
Of the route information stored in the route information storage unit, delete the route information that the user cannot move to the shooting position included in the route information by the shooting time included in the route information. A flow line capturing device comprising: a narrowing means for performing The flow line capturing device according to claim 1,
The narrowing-down means includes, for two pieces of route information stored in the route information storage unit, a photographing time included in one route information and a photographing time included in the other route information. If the difference between the two routes is shorter than the time required to move the distance between the shooting position included in the one route information and the shooting position included in the other route information, the two routes A flow line capturing device, wherein one of the information is deleted. The flow line capturing device according to claim 1 or 2,
A reading device that reads the identification information from a recording medium on which the identification information of the user is recorded;
An interlocking camera that is installed in the vicinity of the reading device and photographs the user when the identification information is read by the reading device; and
The flow line capturing apparatus according to claim 1, wherein the user person image extracted from the image captured by the interlocking camera is used as the user reference person image. The flow line capturing device according to any one of claims 1 to 3,
The verification means includes
Reference feature extraction means for extracting feature information of the user's reference person image;
Person feature extraction means for extracting a person image from images taken by the plurality of cameras and extracting feature information of each of the extracted person images;
The probability that each person image is the person image of the user is obtained from the feature information of the user's reference person image extracted by the reference feature extraction unit and the feature information of the person image extracted by the person feature extraction unit. A flow line capturing apparatus, comprising: accuracy detection means for storing, in the route information storage unit, route information including a photographing time and a photographing position of an image in which a human image having a human image whose accuracy is equal to or greater than a threshold exists. The flow line capturing device according to claim 4,
The accuracy detection unit stores the shooting position and shooting time of the image and the accuracy of the person image as the path information when storing the path information of an image in which a person image having an accuracy equal to or more than a threshold exists.
The narrowing-down means includes, for two pieces of route information stored in the route information storage unit, a photographing time included in one route information and a photographing time included in the other route information. If the difference between the two routes is shorter than the time required to move the distance between the shooting position included in the one route information and the shooting position included in the other route information, the two routes A flow line capturing apparatus, wherein route information having a lower accuracy of information is deleted. The computer collates the reference person image of the user with images taken by a plurality of cameras having different shooting positions, and determines that there is an image showing the user in the images taken by the plurality of cameras. A matching step for storing path information including the shooting time and shooting position of the image in the path information storage unit;
The computer cannot move to the shooting position included in the route information among the route information stored in the route information storage unit by the shooting time included in the route information. And a narrowing step for deleting route information. The flow line capturing method according to claim 6,
In the narrowing-down step, for two pieces of route information stored in the route information storage unit, a shooting time included in one route information and a shooting time included in the other route information If the difference between the two routes is shorter than the time required to move the distance between the shooting position included in the one route information and the shooting position included in the other route information, the two routes A flow line capturing method, wherein one of the information is deleted. The flow line capturing method according to claim 6 or 7,
A reading device that reads the identification information from a recording medium on which the identification information of the user is recorded;
An interlocking camera that is installed in the vicinity of the reading device and that captures the user when the identification information is read by the reading device; and
A flow line capturing method, wherein a person image of the user extracted from an image photographed by the interlocking camera is used as the reference person image of the user. The flow line capturing method according to any one of claims 6 to 8,
The matching step includes
A reference feature extraction step of extracting feature information of the reference person image of the user;
A person feature extraction step of extracting a person image from images taken by the plurality of cameras and extracting feature information of the extracted person images;
From the feature information of the user's reference person image extracted in the reference feature extraction step and the feature information of the person image extracted in the person feature extraction step, the probability that each person image is the person image of the user is determined. A flow line capturing method comprising: an accuracy detection step of storing path information including an imaging time and an imaging position of an image in which a human image having an accuracy of a threshold value or more exists and stored in the path information storage unit. The flow line capturing method according to claim 9, wherein
In the accuracy detection step, when storing route information of an image in which a person image having an accuracy equal to or greater than a threshold value is stored, the shooting position and shooting time of the image and the accuracy of the person image are stored as the route information;
In the narrowing-down step, for two pieces of route information stored in the route information storage unit, a shooting time included in one route information and a shooting time included in the other route information If the difference between the two routes is shorter than the time required to move the distance between the shooting position included in the one route information and the shooting position included in the other route information, the two routes A flow line capturing method, wherein route information having a lower accuracy of information is deleted. A program for causing a computer including a plurality of cameras having different shooting positions and a path information storage unit to function as a flow line capturing device,
The computer,
When the user's reference person image is collated with images taken by the plurality of cameras, and it is determined that there is an image of the user in the images taken by the plurality of cameras, Collating means for storing route information including a photographing time and a photographing position in the route information storage unit;
Of the route information stored in the route information storage unit, delete the route information that the user cannot move to the shooting position included in the route information by the shooting time included in the route information. A program for functioning as a narrowing down means. The program according to claim 11, wherein
The narrowing-down means includes, for two pieces of route information stored in the route information storage unit, a photographing time included in one route information and a photographing time included in the other route information. If the difference between the two routes is shorter than the time required to move the distance between the shooting position included in the one route information and the shooting position included in the other route information, the two routes A program for deleting any one of information. The program according to claim 11 or 12,
The computer is installed in the vicinity of the reading device that reads the identification information from a recording medium on which the user identification information is recorded, and the user is triggered by the reading of the identification information by the reading device. And a linked camera that shoots
A program characterized in that a person image of the user extracted from an image photographed by the interlocking camera is used as the reference person image of the user. The program according to any one of claims 11 to 13,
The verification means includes
Reference feature extraction means for extracting feature information of the user's reference person image;
Person feature extraction means for extracting a person image from images taken by the plurality of cameras and extracting feature information of each of the extracted person images;
The probability that each person image is the person image of the user is obtained from the feature information of the user's reference person image extracted by the reference feature extraction unit and the feature information of each person image extracted by the person feature extraction unit. A program comprising: accuracy detection means for storing route information including a photographing time and a photographing position of an image in which a human image having a certain accuracy or more than a threshold value exists is stored in the route information storage unit. The program according to claim 14, wherein
The accuracy detection unit stores the shooting position and shooting time of the image and the accuracy of the person image as the path information when storing the path information of an image in which a person image having an accuracy equal to or more than a threshold exists.
The narrowing-down means includes, for two pieces of route information stored in the route information storage unit, a photographing time included in one route information and a photographing time included in the other route information. If the difference between the two routes is shorter than the time required to move the distance between the shooting position included in the one route information and the shooting position included in the other route information, the two routes A program characterized by deleting route information with a lower accuracy of information.

Images