JP2016014936A - Movement pathway identification device - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide means enabling the movement pathway of a person to be identified.SOLUTION: A movement pathway identification system 1 includes a plurality of automatic ticket gates 11 arranged at ticket gates of a station, imaging units 12 provided according to the automatic ticket gates 11 and photographing the face of a passerby passing through the automatic ticket gates 11, and a movement pathway identification device 13 which identifies a person having passed through one automatic ticket gate in the past and a person newly passing through one automatic ticket gate 11 on the basis of the face image photographed by the imaging unit 12, and which groups photographing point passing data showing information on passing of a photographing point (a position with the imaging unit 12 arranged) of the same person and manages the grouped data. The movement pathway identification device 13 collates face feature data generated from face image data transmitted, for example, from a terminal device 15 used by a searching person searching for a specific person with face feature data of a stored person, and transmitting movement pathway data showing the movement pathway of a person determined to be the same person to the terminal device 15.

Description

  The present invention relates to a technique for specifying a travel route of a passerby.

  A mechanism for preventing unauthorized use of a pass ticket used for passing a ticket gate has been proposed. For example, in Patent Document 1, a face image is recorded on an IC card, and when the IC card is used for traffic in an automatic ticket gate or the like, a face image of a passer-by is compared with a face image read from the IC card. A mechanism for preventing unauthorized use of an IC card by a third party has been proposed.

  Patent Document 2 records the issuance information encrypted based on the purchaser's face image data at the time of issuance of a pass ticket in a pass ticket, and when the pass ticket is used in an automatic ticket gate, There has been proposed a mechanism for preventing unauthorized use of a pass ticket by a third party by decoding issued information read from the pass ticket based on face image data.

  Also, in Patent Document 3, a facial image of a user and a flag indicating whether or not there is a fee discount are registered in a server in association with the ID of a wireless card used by a paying facility user for entry, and the wireless card is automatically When used in the ticket gate, the ID and passer's face image received from the wireless card are transmitted to the server by the automatic ticket gate, and the face image associated with the ID and the face image transmitted from the ticket gate By sending the verification result and flag to the automatic ticket checker, it is possible to prevent unauthorized use of the wireless card by a third party, and when the wireless card is used legitimately, a system is proposed to discount the fee according to the flag. ing.

  Further, in Patent Document 4, each of a plurality of automatic ticket gates for entry arranged at a station uses a pass ticket identification code and a user image used by a user who has performed ticket gate processing via a server. When a pass ticket with the same identification code is used more than once, a mechanism that prevents the use of a pass ticket by comparing the image of a user who has participated in the past with the image of a user who has newly participated Has been proposed.

JP 2012-128527 A JP 2012-79234 A JP 2010-97272 A JP 2009-205352 A

  According to the mechanism proposed in Patent Documents 1 to 4 described above, unauthorized use is immediately detected at the timing when the pass ticket is used in an automatic ticket gate or the like, and an unauthorized passerby is detected when unauthorized use is detected. Is not permitted.

  For example, an automatic ticket checker at a station at rush hour is required to perform a ticket check process as quickly as possible so as not to cause a traffic jam at the ticket gate. For this reason, it is not practical to immediately determine passers-by for passersby and allow or prohibit them based on face image matching. For example, there is a misjudgment in face image matching, which prevents a passerby from passing a legitimate person, resulting in a drop in service for the passerby and a traffic jam at the ticket gate, and for other passers-by. This is because the service also deteriorates.

  In recent years, in addition to magnetic tickets, IC cards using short-range wireless communication have been used as passing tickets. A passerby who uses an IC card using short-range wireless communication may pass through an automatic ticket gate without noticing that when reading of ticket information from the IC card fails.

  Therefore, when there are passers-by who repeat the illegal use of the pass ticket, it is necessary to identify such an illegal passer-by and take measures such as reporting to the police depending on the situation. In this case, if the passerby's past movement route is known, it becomes an important clue when the police or the like searches for an illegal passerby.

  The above example is a case of capturing a fraudulent person, but the same applies when searching for a missing person, for example.

  In view of the above circumstances, an object of the present invention is to provide a means for enabling a person's movement route to be specified.

  In order to solve the above-described problem, the present invention provides shooting point passage data indicating at least the characteristics of the face of a person shot at each of a plurality of shooting points, the shooting time, and the shooting location. Grouping means for grouping each of the facial features indicated by the shooting point passage data indicating the same person and storing them in the storage device, and the features indicated by the shooting point passage data based on the plurality of grouped shooting point passage data An apparatus including a specifying unit that specifies a movement path of a person on the face is provided as a first aspect.

  According to the apparatus of the first aspect, it is possible to specify the movement path of the person who has passed the shooting point.

  In the apparatus of the first aspect, each of the plurality of shooting points is an entrance / exit, the shooting point passage data indicates whether entry or exit, and among the shooting point passage data grouped in the same group, Based on the time indicated by the shooting point passage data and the entry and exit, the pairing means for pairing the shooting point passage data indicating entry and participation as a pair is provided, and the specifying means includes the paired shooting point passage data. A configuration in which the movement path of the person is specified based on the above may be adopted as the second aspect.

  According to the apparatus of the 2nd aspect, the entrance / exit with respect to the facility of the person who passed the imaging | photography point can be specified.

  In the apparatus of the second aspect, each of the plurality of shooting points is a ticket gate of a station, and when the grouping unit groups shooting point passage data indicating participation, the time and place indicated by the shooting point passage data A configuration in which group candidates of grouping destinations of the photographing point passage data are specified based on the time indicating the time required for moving between stations may be employed as the third aspect.

  According to the apparatus of the third aspect, when the facility where the person enters and exits is a transportation facility, the processing for specifying the movement route is reduced by using the time required for movement between stations.

  In any one of the first to third devices, the shooting point passage data indicates identification information of a ticket used by a person to pass the shooting point, and the grouping means performs the shooting point passage when grouping the shooting point passage data. As a fourth aspect, a configuration in which a group including shooting point passage data indicating identification information of the same ticket as identification information of the ticket indicated by the shooting point passage data is specified as a group candidate of the passing data grouping destination is a fourth aspect. It may be adopted.

  According to the apparatus of the fourth aspect, when a person uses a voucher on which identification information is recorded when passing a shooting point, the process for specifying the movement route is reduced by using the voucher identification information.

  In any one of the first to fourth apparatuses, a part of the shooting point passage data indicates identification information of a ticket used by a person to pass the shooting point, and the grouping unit is configured to group the shooting point passage data. When the shooting point passage data indicates ticket identification information, a group including shooting point passage data indicating the same ticket identification information as the ticket identification information as a group candidate of the shooting point passage data grouping destination A configuration may be adopted as the fifth aspect in which grouping based on the similarity of facial features is performed using a threshold value lower than that in the case where the shooting point passage data does not indicate ticket identification information after specifying .

  According to the apparatus of the 5th aspect, the misjudgment in identification of the passerby who passed the imaging | photography point using the same ticket is reduced.

  In any one of the first to fifth devices, when grouping the shooting point passage data, the grouping unit creates a new group including the shooting point passage data when there is no grouping destination group of the shooting point passage data. The configuration of generating the data may be adopted as the sixth aspect.

  According to the device of the sixth aspect, pre-registration of the passer's facial features is not required.

  In any one of the first to sixth devices, the grouping unit sets a predetermined condition that an elapsed time from the latest time among the times indicated by the shooting point passage data included in the group is a predetermined condition. If the condition is satisfied, a configuration of causing the storage device to delete the data of the group may be employed as the seventh aspect.

  According to the device of the seventh aspect, it is possible to reduce the storage capacity required for the storage device by deleting the information of the person whose time has passed to such an extent that identification based on facial features is impossible.

The figure which showed the whole structure of the movement path | route identification system concerning one Embodiment. The figure which showed the function structure of each apparatus which comprises the movement path | route identification system concerning one Embodiment. The figure which illustrated the data composition of required time DB concerning one embodiment. The figure which illustrated the data composition of log DB concerning one embodiment. The figure which illustrated the data composition of person DB concerning one embodiment. The figure which illustrated the data composition of passage history DB concerning one embodiment. The figure which illustrated the data composition of inconsistency log DB concerning one embodiment. The flowchart of the process which the grouping part of the movement path | route identification apparatus concerning one Embodiment performs (when ticket ID is unusable and it is passage of admission). The flowchart of the process which the grouping part of the movement path | route identification apparatus concerning one Embodiment performs (when ticket ID is unusable and it is passage of participation). The flowchart of the process which the grouping part of the movement path | route identification apparatus concerning one Embodiment performs (when ticket ID is available and it is passage of admission). The flowchart of the process which the grouping part of the movement path | route specific | specification apparatus concerning one Embodiment performs (when ticket ID can be utilized and it is passage of participation). The figure which illustrated the screen displayed on the terminal unit concerning one embodiment.

[Embodiment]
The movement path | route identification system 1 concerning one Embodiment of this invention is demonstrated below. FIG. 1 is a diagram showing the overall configuration of the movement path specifying system 1. The movement route specifying system 1 includes a plurality of automatic ticket gates 11 (1) to 11 (i) (where i is a natural number indicating the total number of automatic ticket gates) arranged at the entrance or exit of a station, and automatic ticket gates. Imaging devices 12 (1) to 12 (i) that are provided corresponding to each of the devices 11 (1) to 11 (i) and photograph a face of a passerby that passes through a corresponding automatic ticket gate (an example of a shooting point). ), A movement path specifying device 13 for specifying a movement path of the passer-by based on the face image data indicating the face image of the passer-by photographed by each of the imaging devices 12 (1) to 12 (i), and a movement path A storage device 14 that stores various data used by the specific device 13 and a terminal device 15 that is operated by a user who inquires the movement route specifying device 13 about a specific person's movement route are provided.

  Hereinafter, the automatic ticket gates 11 (1) to (i) are collectively referred to as the automatic ticket gate 11, and the image pickup devices 12 (1) to 12 (i) are collectively referred to as the image pickup device 12. Although the number of terminal devices 15 shown in FIG. 1 is one, the number of terminal devices 15 is arbitrarily changed according to the number of users who inquire about the movement route of a person. In FIG. 1, the movement route identification device 13 is shown as one device, but the movement route identification device 13 may be configured by a plurality of device groups operating in association with each other. Further, FIG. 1 shows a configuration in which the storage device 14 is directly connected to the movement route specifying device 13, but the storage device 14 exchanges data with the movement route specifying device 13 via the network. A configuration may be employed.

  The automatic ticket checker 11 and the image pickup device 12 are connected so that data communication can be performed between devices corresponding to each other by wire or wirelessly. Further, each of the automatic ticket gates 11 and the movement route specifying device 13 can communicate data with each other via the network 9. Further, the terminal device 15 and the movement path specifying device 13 can perform data communication with each other via the network 9.

  The automatic ticket gate 11 has the same configuration as that of a normal automatic ticket gate except that it includes a communication unit that performs data communication with the corresponding imaging device 12.

  The image pickup device 12 has the same configuration as that of a normal image pickup device except that the image pickup device 12 includes a communication unit that performs data communication with the corresponding automatic ticket gate 11.

  The movement path identification device 13 is realized by a general-purpose server computer executing processing according to the program for the movement path identification device 13 according to the present embodiment. That is, the computer used for realizing the movement path specifying device 13 includes a memory, a processor, and a communication interface, and the processor executes a process according to a program read from the memory, thereby having a functional configuration described later. 13 functions. However, the movement path specifying device 13 is not limited to a configuration realized by a general-purpose computer, and may be configured as a dedicated device using a logic integrated circuit such as an ASIC.

  The storage device 14 is a large-capacity storage device in which nonvolatile memories such as HDD and SDD are integrated.

  The terminal device 15 is realized by a general-purpose terminal device computer executing processing according to the program for the terminal device 15 according to the present embodiment. That is, the computer used to implement the terminal device 15 includes a memory, a processor, a communication interface, an operation device such as a keyboard, a display device such as a liquid crystal display, and the processor executes processing according to a program read from the memory, It functions as a terminal device 15 having a functional configuration described later. However, the terminal device 15 is not limited to a configuration realized by a general-purpose computer, and may be configured as a dedicated device using a logic integrated circuit such as an ASIC.

  FIG. 2 is a diagram showing a functional configuration of each device constituting the movement path specifying system 1. The automatic ticket gate 11 includes a passage sensor 111 that detects passage of a passerby, a transmission unit 112 that transmits photographing instruction data indicating a photographing instruction to the image pickup device 12 to the image pickup device 12 according to a detection result of the passage sensor 111, and a shooting instruction. A reception unit 113 for receiving face image data indicating a passerby's face image transmitted from the image pickup device 12 as a response to the data transmission, and a current time indicating the current time by continuously counting the elapsed time from the reference time The time counting unit 114 that generates time data, the generation unit 115 that generates shooting point passage data (described later) triggered by reception of face image data by the reception unit 113, and the shooting point passage data generated by the generation unit 115 are moved. A transmission unit 116 that transmits to the route specifying device 13 and a reading unit 117 that reads information from a ticket used by a passer-by when the automatic ticket gate 11 passes.

  The imaging device 12 receives imaging instruction data transmitted from the automatic ticket gate 11 and imaging that captures a passerby's face and generates face image data according to the imaging instruction data received by the receiving unit 121. And a transmission unit 123 that transmits the face image data generated by the imaging unit 122 to the automatic ticket gate 11.

  The storage device 14 stores various types of data used by the movement path identification device 13 in five databases: a required time DB (Database), a log DB, a person DB, a passage history DB, and an inconsistency log DB.

  The required time DB stores required time data indicating a standard required time when traveling between any two stations among a plurality of stations where any of the automatic ticket gates 11 are arranged by means of transportation. It is a database.

  FIG. 3 is a diagram illustrating a data configuration of the required time DB. The required time DB is a collection of records for each combination of two stations, and each record has fields [departure station], [target station], and “required time.” The field [departure station] identifies the departure station. Information (for example, station name) is stored, and identification information (for example, station name) of the target station is stored in the field [target station], and the required time for moving from the departure station to the target station is stored in the field [required time]. Is stored.

  The log DB is a database that accumulates information about the passage when any passer-by passes through any of the automatic ticket gates 11.

  FIG. 4 is a diagram illustrating a data configuration of the log DB. The log DB is a collection of records relating to each passage, and each record has fields [passing ID], [passing time], [passing place], [entrance / exit], [face feature], [ticket ID], [ticket]. Type]. In the field [Pass ID], a pass ID for identifying each pass is stored. In the field [passing time], time data indicating the current time at the time of passing is stored.

  In the field [passing place], place data for identifying the place where the automatic ticket gate 11 (that is, the image pickup device 12) is arranged is stored. The location data is, for example, data such as “XX station ×× ticket gate ΔΔ ticket gate”. In the field [entrance / exit], data indicating whether the automatic ticket gate 11 is for entry or entry is stored. Data stored in the field [entrance / exit] is data such as “entrance” and “entrance”, for example.

  In the field [ticket ID], a ticket ID for identifying a ticket used by a passer-by for passing the automatic ticket gate 11 is stored. In the field [ticket type], data indicating the type of ticket used by the passer-by for the automatic ticket gate 11 is stored. The ticket type is, for example, data such as “magnetic ticket”, “IC card (user limited)”, “IC card (user limited)”.

  The person DB is a database that accumulates information about a person who has passed through any of the automatic ticket gates 11 in the past.

  FIG. 5 is a diagram illustrating a data configuration of the person DB. The person DB is a collection of records for each person, and each record has fields [person ID], [face feature], [update time], [ticket ID], and [ticket type]. The field [person ID] stores a person ID for identifying each person identified by the facial feature. In the field [face feature], face feature data indicating the feature amount of the face indicated by the face image data generated by the imaging device 12 is stored. The field [update time] stores time data indicating the time when the facial feature data was last updated.

  In the field [ticket ID], a ticket ID of a ticket of a type that can be used repeatedly such as an IC card used by a person in the past is stored. Since the same person may use a plurality of tickets, a plurality of ticket IDs may be stored in the field [ticket ID]. In the field [ticket type], data indicating a ticket type corresponding to each ticket ID is stored. Therefore, if the number of ticket IDs stored in the field [ticket ID] is plural, the same number of data as the number of ticket IDs is stored in the field [ticket type].

  The passage history DB is a database that accumulates, for each person, a history of passages that the person has made in the past in chronological order.

  FIG. 6 is a diagram illustrating a data configuration of the passage history DB. The passage history DB is a collection of tables for each person identified by the person ID, and each table is a collection of records for each passage identified by the passage ID. Each record has fields [passing ID], [passing time], [passing place], [entrance / exit], [ticket ID], [ticket type], and [pair ID]. Each of the fields [passing ID], [passing time], [passing place], [entrance / exit], [ticket ID], and [ticket type] included in the table of the passing history DB has a log DB (FIG. 4). Data similar to the data stored in the field with the same name is stored.

  In the field [Pair ID], a passage ID for identifying the passage of a partner to be a pair of entrance and exit is stored. For example, after the target person passes through the automatic ticket gate 11 (k) of the X station, which is the departure station, enters the station premises and moves using transportation, the automatic ticket gate of the Y station, which is the target station. 11 (l), when entering from the station premises, the record field [Pair ID] relating to entry at X station stores the passage ID of entry at Y station, and the field of record relating to entry at Y station [ The pair ID] stores the passage ID for the entrance to the X station.

  The inconsistency log DB is a database that accumulates data relating to passages in which a partner of an entrance / exit could not be identified among the passages that the person has made in the past for each person.

  FIG. 7 is a diagram illustrating a data configuration of the inconsistency log DB. Data stored in the inconsistency log DB is data of a record whose field [pair ID] is empty data among records stored in the passage history DB (FIG. 6). Accordingly, the fields included in the record of the inconsistency log DB are all fields except the field [pair ID] among the fields included in the record of the passage history DB.

  Returning to FIG. 2, the description of the functional configuration of each device constituting the movement path specifying system 1 will be continued. The movement path specifying device 13 receives the shooting point passage data transmitted from the automatic ticket gate 11 and the request data (described later) transmitted from the terminal device 15, and the photographing received by the receiving unit 131. The face feature specifying unit 132 that generates the face feature data indicating the specified feature amount by specifying the face feature amount indicated by the face image data included in the point pass data, and the face feature data specified by the face feature specifying unit 132 Based on the identification of the person, the grouping unit 133 groups the shooting point passage data for each person who has passed, and the shooting point passage data grouped by the grouping unit 133 has an entry and participation relationship. A pairing unit 134 for pairing photographing point passage data indicating passage is provided.

  Further, when the reception unit 131 receives the request data transmitted from the terminal device 15, the movement path identification device 13 identifies the movement path of the person designated according to the request data and indicates the identified movement path. A travel route specifying unit 135 that generates travel route data and a transmission unit 136 that transmits the travel route data generated by the travel route specifying unit 135 to the requesting terminal device 15 are provided.

The terminal device 15 includes an operation unit 151 that receives a user operation, a transmission unit 152 that transmits request data for requesting movement route data to the operation unit 151 according to an operation performed by the user, to the movement route identification device 13. The receiving unit 153 that receives the moving route data transmitted from the moving route specifying device 13 as a response to the request data transmitted from the transmitting unit 152, and the moving route indicated by the moving route data received by the receiving unit 153 is the user. Is provided with a display unit 154.

  Below, operation | movement of the movement path | route identification system 1 is demonstrated. First, each time the passerby passes, each of the automatic ticket gates 11 transmits shooting instruction data to the imaging device 12 by the transmission unit 112 when the passage sensor 111 detects the passage. When the imaging unit 12 receives the shooting instruction data transmitted from the automatic ticket gate 11 by the receiving unit 121, the imaging unit 122 captures the face of the passerby and the generated face image data is transmitted to the automatic ticket gate 11 by the transmission unit 123. Send to.

  When the automatic ticket checker 11 receives the face image data transmitted from the image pickup device 12 by the reception unit 113, the generation unit 115 generates photographing point passing data including the face image data. Compared to the data structure of the record in the log DB (FIG. 4), the photographing point passage data is stored in the face image data in place of the point [field ID] and the field [face feature]. It is the same except that it has a field [face image]. That is, the shooting point passage data has fields [passing time], [passing place], [entrance / exit], [face image], [ticket ID], and [ticket type].

  In the field [passing time] of the shooting point passage data, the current time data generated by the timer unit 114 at the time when the shooting point passage data is generated by the generation unit 115 is stored. In the fields [passing place] and [entrance / exit], data stored in advance in the automatic ticket gate 11 is stored. The field [face image] stores the face image data received from the image pickup device 12 as described above. In the fields [ticket ID] and [ticket type], data read from the ticket held by the passerby by the automatic ticket gate 11 is stored.

  The automatic ticket gate 11 transmits the shooting point passage data generated by the generation unit 115 to the movement path identification device 13 by the transmission unit 116. When the moving part specifying device 13 receives the shooting point passage data transmitted from the automatic ticket gate 11 by the receiving unit 131, the movement path specifying device 13 determines the facial feature amount indicated by the face image data included in the received shooting point passage data. Face feature data that is specified by the specifying unit 132 and that indicates the specified feature amount is generated. The face feature identification unit 132 adds a new record to the log DB (FIG. 4), generates a unique passing ID according to a predetermined rule, stores it in the field [passing ID] of the new record, and from the automatic ticket gate 11 The received field data other than the field [face image] is stored in the field with the same name in the new record. Further, the face feature specifying unit 132 stores the generated face feature data in the field [face feature] of the new record.

  Note that the face image taken by the image pickup device 12 may be low quality so that a person cannot be identified. For example, when shooting is performed in a state where a passer's face is blocked by a hat or a hand, the facial feature quantity cannot be specified from the face image. In this case, the field [face feature] of the new record in the log DB is empty data.

  When a new record is added to the log DB, the grouping unit 133 performs processing for grouping the passage indicated by the newly added record for each person who has made the passage. Hereinafter, an outline of the grouping process performed by the grouping unit 133 will be described with reference to a flowchart.

  When the field [face feature] of the record newly added to the log DB is not empty data, the grouping unit 133 passes based on the data stored in the fields [entrance / exit] and [ticket ID]. (1) If the ticket ID is not available and the entrance is passing, (2) If the ticket ID is not available and the entry is passing, (3) If the ticket ID is available and the entrance is passing, 4) When the ticket ID is available and the entry is passing, it is determined which of the four is, and processing is performed according to a flow corresponding to each of these.

  Note that when the ticket ID is not usable, for example, when a passerby holds the IC card that allows the ticket information to be read by short-range wireless communication over the automatic ticket checker 11, the reading unit 117 of the automatic ticket checker 11 This is the case where the ticket information has failed to be read but the passer-by has passed as it is, or the passer-by has passed through the automatic ticket gate 11 due to a magnetic ticket having no ticket ID. In these cases, the field [ticket ID] of the record in the log DB is empty data.

  (1) When the ticket ID is not available and the entry is passing, the grouping unit 133 performs processing according to the flow shown in FIG. First, the grouping unit 133 stores the face feature data stored in the field [face feature] of the record newly added to the log DB (FIG. 4) and the fields [ The facial feature data stored in “Facial Features” are collated, and the similarity of the facial features indicated by these facial feature data is calculated (S101). The grouping unit 133 determines whether or not the person with the facial feature indicated by the facial feature data is the same person by determining whether or not the highest similarity among the calculated similarities is higher than a predetermined threshold T1. It is determined whether or not (S102).

  Hereinafter, the determination of whether or not the person with the face feature indicated by the two face feature data is the same person is referred to as “face feature data collation” for convenience, and it is determined that these persons are the same person. The face feature data has been successfully verified. In other words, the collation of the face feature data means not the collation of whether the face feature data itself is the same, but the identification of the person having the face feature indicated by the face feature data.

  When the face feature data collation fails in S101 (S102; “No”), the grouping unit 133 newly creates a person whose face feature indicated by the face feature data of the record newly registered in the log DB is not registered in the person DB. A record relating to this person is newly registered in the person DB, and a new table corresponding to the person ID of the person newly registered in the passage history DB (FIG. 6) is added (S103). That is, the grouping unit 133 adds a new record to the person DB, stores it in the field [person ID] of the new record to which a unique person ID determined by a predetermined rule is added, and stores the target record in the log DB. The data of the field [face feature] and [passing time] is stored in the field [face feature] and [update time] of the new record of the person DB. Then, a table corresponding to the person ID of the new record in the person DB is added to the passage history DB. The process of S103 is a process of generating a group related to a new person.

  When the facial feature data is successfully collated in S101 (S102; “Yes”), the grouping unit 133 indicates that the facial feature person indicated by the facial feature data in the record newly registered in the log DB is the person in the person DB that has been successfully collated. It is determined that the person corresponds to the record, and the record of the person DB related to this person is updated (S104). That is, the grouping unit 133 overwrites and updates the data of the fields [face features] and [update time] of the record of the person DB that has been successfully matched with the fields [face features] and [pass times] of the target records of the log DB. To do.

  The grouping unit 133 updates the passage history DB (FIG. 6) after the process of S103 or S104 (S105). That is, the grouping unit 133 adds the same name of the new record in the log DB to each of the fields in the table of the passage history DB corresponding to the person ID of the record newly added to the person DB in S103 or the person ID of the record updated in S104. Stores field data. The process of S105 is a process of grouping new shooting point passage data related to the person for the group for each person.

  (2) When the ticket ID is unavailable and the entry is passing, the grouping unit 133 performs processing according to the flow shown in FIG. First, the grouping unit 133 uses the required time data stored in the required time DB (FIG. 3), and uses the required time data stored in the required time DB (FIG. 3) as a candidate for a person who is the same as the person corresponding to the record newly registered in the log DB. (Referred to as “person candidates”) (S202). Specifically, the grouping unit 133 extracts a table satisfying the following conditions from the passage history DB (FIG. 6), and specifies the person ID of the extracted table as the person ID of the same person candidate.

The field [entrance / exit] of the last record in the passage history DB table is “entrance”.
The field of the record of the required time DB with the station indicated by the field [passing place] of the last record in the passing history DB as the departure station and the destination station indicated by the field [passing place] of the new record DB of the log DB [ The difference between the time indicated by the “required time” and the time indicated by the field [passing time] of the last record in the table of the passage history DB until the time indicated by the field [passing time] of the new record of the log DB is a predetermined time. Is within.

  Subsequently, the grouping unit 133 determines whether the same person candidate has been successfully identified (S202). In S201, when no person ID is specified as the person ID of the same candidate (S202; “No”), the grouping unit 133 moves the process to S101 (FIG. 8).

  On the other hand, when any person ID is identified as the person ID of the same person candidate in S201 (S202; “Yes”), the grouping unit 133 determines the facial feature data of each identified record of the same person candidate, The face feature data of the new record in the log DB is collated (S203). Specifically, the grouping unit 133 calculates the similarity of the facial features indicated by the two facial feature data, and determines whether or not the highest similarity among the calculated similarities is higher than a predetermined threshold T1. It is determined whether or not the persons having the characteristics indicated by the two face characteristic data are the same person. The threshold T1 used in S203 is the same threshold as the threshold T1 used in S101.

  The grouping unit 133 determines whether or not the collation in S203 is successful (S204). When the facial feature data is successfully verified in S203 (S204; “Yes”), the grouping unit 133 moves the process to S104 (FIG. 8). On the other hand, when the face feature data collation fails in S203 (S204; “No”), the grouping unit 133 moves the process to S101 (FIG. 8).

  (3) When the ticket ID is available and the entry is passing, the grouping unit 133 performs processing according to the flow shown in FIG. First, the grouping unit 133 extracts the same person candidate using the ticket ID stored in the person DB (FIG. 5) (S301). Specifically, the grouping unit 133 extracts a record in which the field [ticket ID] matches the field [ticket ID] of the new record in the log DB from the person DB, and sets the person ID of the extracted record as a person with the same person candidate. Specify as ID.

  Subsequently, the grouping unit 133 determines whether the identification of the same person candidate has succeeded (S302). If no person ID is identified as the person ID of the same person candidate in S301 (S302; “No”), the grouping unit 133 moves the process to S101 (FIG. 8).

  On the other hand, if any person ID is identified as the person ID of the same person candidate in S301 (S302; “Yes”), the grouping unit 133 determines the facial feature data of each identified record of the same person candidate, The face feature data of the new record in the log DB is collated (S303). Specifically, the grouping unit 133 calculates the similarity of the facial features indicated by the two facial feature data, and determines whether the highest similarity among the calculated similarities is higher than a predetermined threshold T2. It is determined whether or not the persons having the characteristics indicated by the two face characteristic data are the same person.

  The threshold T2 used in S303 is a threshold lower than the threshold T1 used in S101 and S203. The reason why the low threshold value is used in S303 is that the matching of the facial feature data in S303 is identification between persons using the same ticket. Therefore, there is a high possibility that those persons are the same person. In order to reduce the risk of misjudgment due to changes in facial features of the same person due to the presence or absence of hairstyle or wrinkles (misjudgment to judge the same person as another person) This is because it is desirable to lower the identification standard. However, in consideration of the former risk and the latter risk, the threshold value T1 used in S101 and S203 may be used in S303.

  The grouping unit 133 determines whether or not the collation in S303 is successful (S304). When the facial feature data is successfully collated in S303 (S304; “Yes”), the grouping unit 133 moves the process to S104 (FIG. 8). In this case, in the update of the person DB, when data is already stored in the fields [ticket ID] and [ticket type], the grouping unit 133 additionally updates these data instead of overwriting.

  On the other hand, when the facial feature data verification fails in S303 (S304; “No”), the grouping unit 133 shifts the processing to S101 (FIG. 8).

  (4) When the ticket ID is available and the entry is passing, the grouping unit 133 performs processing according to the flow shown in FIG. The process of S301 to S304 shown in FIG. 11 is the same as the process of S301 to S304 shown in FIG. However, when the ticket ID is available and it is a pass of participation, in the determination in S302, when it is determined that the extraction of the same person candidate using the ticket ID in S301 has failed (S302; “No”), the grouping unit In step 133, the process proceeds to step S201 (FIG. 9) instead of step S101. Further, when the ticket ID is available and the entry has passed, if it is determined in S304 that the facial feature data collation in S302 has failed (S304; “No”), the grouping unit 133 performs the process in S101. Instead, the process proceeds to S201 (FIG. 9).

  The above is the processing executed by the grouping unit 133 when a new record is added to the log DB.

  The pairing unit 134 enters and exits a pair of records (shooting point passage data) stored in the passage history DB (FIG. 6) at a timing determined according to a predetermined rule (for example, every elapse of a predetermined time). A pairing process is performed for every two records shown.

  Specifically, the pairing unit 134, in each table of the passage history DB, for each record in which the field [pair ID] is empty data, the preceding and following records (the field [pair ID] is empty data) If there is a combination of two records satisfying the following condition, the record is identified as a pair of records.

  The station indicated by the field [passing place] of the record whose field [entrance / exit] is “entrance” is the departure station, and the station indicated by the field [passing place] of the record whose field [entry / exit] is “participation” is the target station. From the time indicated by the field [Required time] of the record in the required time DB (FIG. 3) and the time indicated by the field [passing time] of the record whose field [Entry / Exit] is “Entry”, the field [Entry / Exit] The difference from the time until the time indicated by the field [passing time] of the record that is “participation” is within a predetermined time.

  When the pairing is successful, the pairing unit 134 stores the passing ID of each pair of counterparts in the field [Pair ID] of the corresponding two records of the passing history DB.

  Further, the pairing unit 134 updates the passage history DB in the pairing process and also updates the inconsistency log DB (FIG. 7). Specifically, the pairing unit 134 extracts a record in which [Pair ID] is empty data from the records in each table stored in the updated passage history DB, and uses the extracted record to create the inconsistency log DB. Overwrite and update.

  Based on the photographing point passage data transmitted from the automatic ticket gate 11 by the processing performed by the grouping unit 133 and the pairing unit 134, a log DB (FIG. 4), a person DB (FIG. 5), a passage history DB (FIG. 6) The inconsistency log DB (FIG. 7) is continuously updated.

  The terminal device 15 can inquire the movement route of the specific person to the movement route identification device 13 at an arbitrary timing. FIG. 12 is a diagram illustrating a screen displayed on the display unit 154 in response to a user operation of the terminal device 15. The user of the terminal device 15 operates the operation unit 151 of the terminal device 15 to cause the display unit 154 to display the face image designation screen illustrated in FIG.

  On the face image designation screen, after the user performs a click operation on the “Open” button, etc., the user selects the face image data indicating the face image of the person whose movement route is to be queried, and clicks on the “Inquire” button. When an operation or the like is performed, the terminal device 15 generates request data that includes the face image data designated by the user and requests transmission of movement route data indicating the movement route, and the generated request data is transmitted to the movement route by the transmission unit 152. Transmit to the specific device 13.

  When the movement path specifying device 13 receives the request data by the receiving unit 131, the movement path data indicating the movement path of the person of the face image indicated by the face image data included in the request data by the face feature specifying unit 132 and the movement path specifying unit 135. Is generated. Specifically, first, the face feature specifying unit 132 specifies the feature value of the face indicated by the face image data received from the terminal device 15, and generates face feature data indicating the specified feature value. Subsequently, the movement path specifying unit 135 compares the face feature data generated by the face feature specifying unit 132 with the face feature data stored in each field [face feature] of the record of the person DB (FIG. 5). I do. When the facial feature data is successfully verified, the face feature specifying unit 132 reads a table corresponding to the person ID of the person identified as the same person from the passage history DB (FIG. 6), ], The data stored in [passing time] is associated with each record, and data obtained by sorting by the time indicated by the data stored in the field [passing time] is generated as travel route data.

  The movement path specifying device 13 transmits the movement path data generated as described above to the requesting terminal device 15 by the transmission unit 136. In addition, when the collation of the facial feature data by the movement route identification unit 135 fails, the movement route identification device 13 sends response data indicating “no applicable person” to the request source terminal device 15 instead of the movement route data. Send.

  When the terminal device 15 receives the travel route data or the response data transmitted from the travel route specifying device 13 by the receiving unit 153, the terminal device 15 displays a screen showing the inquiry result on the display unit 154 according to the received data. FIG. 12B is a diagram illustrating a travel route display screen displayed on the display unit 154 when the terminal device 15 receives travel route data from the travel route specifying device 13. On the movement path display screen, past movement paths of the person of the face image specified by the user are displayed by a list and a plot on the map. In the map of the moving route display screen illustrated in FIG. 12, the movement between two shooting points adjacent to each other at the passage time when the entrance / exit is paired is indicated by a solid line arrow, and the passage time that is not paired The movement between two shooting points adjacent to each other is indicated by a dashed arrow. Note that the display mode of the movement route is not limited to that illustrated in FIG.

  According to the movement route specifying system 1 described above, since the movement route of the person to be searched is promptly provided to the person who searches for the person, it becomes easy to search for the person.

  Since the log DB, person DB, passage history DB, and inconsistency log DB are continuously updated and new records are added, grouping is performed so that the storage capacity of the storage device 14 for storing these data does not become insufficient. The unit 133 instructs the storage device 14 to delete a record satisfying a predetermined condition among records stored in these databases at a timing determined according to a predetermined rule (for example, every elapse of a predetermined time).

  For example, the grouping unit 133 records, among the records in the person DB, the field [update time] that has passed a predetermined time from the current time, the passage history DB and the inconsistency log corresponding to the person ID of the record. The storage device 14 is made to delete the DB table. In this case, as the predetermined time, for example, a time (for example, 10 years) is employed such that the person cannot be identified using the old face feature data as a result of the facial features changing with age. It is desirable that

  In addition, the grouping unit 133 causes the storage device 14 to quickly delete a record relating to the shooting point passage data stored in the passage history DB, for example, among the records in the log DB. This is because these records have the same information stored in the passage history DB, so that even if they are deleted, there is no inconvenience.

  Further, the grouping unit 133 causes the storage device 14 to delete a record whose field [passing time] has exceeded a predetermined time from the current time among the records of the passing history DB and the inconsistency log DB, for example. In this case, as the predetermined time, for example, it is desirable to adopt a time (for example, one year) to the extent that old movement route data is not useful for searching for a person as a result of a change in the movement route pattern of the person.

Note that the data deletion instruction rule by the grouping unit 133 is an example, and the data deletion instruction may be performed according to other rules.
[Modification]
The embodiment described above can be variously modified. Examples of these modifications are shown below. Note that the above-described embodiment and the modifications described below may be combined as appropriate.

[First Modification]
In the description of the above-described embodiment, the data stored in the log DB, person DB, passage history DB, and mismatch log DB managed by the movement path specifying device 13 are used for searching for people. The purpose of using this data is not limited to searching for people. For example, the data stored in the inconsistency log DB may be used for identifying a person who illegally uses a ticket, identifying a problem in a ticket gate system that allows unauthorized use, and the like.

  Further, the same ticket ID is stored in the field [ticket ID] of a plurality of records stored in the person DB, and the data of the field [ticket type] corresponding to these ticket IDs is, for example, “magnetic ticket” or “ In the case of “IC card (user only)”, there is a possibility that a plurality of persons identified by the person IDs of these records are illegally using the same IC card. Therefore, these data stored in the person DB can also be used for specifying the person who is illegally using the ticket.

[Second Modification]
In the embodiment described above, the process of generating face feature data from the face image data is performed by the face feature specifying unit 132 of the movement path specifying device 13. The place where the process of generating the facial feature data from the facial image data is performed is not limited to the movement path specifying device 13, and may be performed by, for example, the imaging device 12 that generates facial image data, or the facial image data from the imaging device 12. May be performed in the automatic ticket checker 11 that receives.

[Third Modification]
In the above-described embodiment, the time data indicating the passage time included in the photographing point passage data is time data generated by the timer 114 of the automatic ticket gate 11. Instead of this, for example, the moving route specifying device 13 includes a timekeeping unit, and time data indicating the current time when the moving route specifying device 13 receives the shooting point passage data from the automatic ticket gate 11 is used as the shooting point passage data. The structure used as the passage time of passage indicated by may be adopted.

[Fourth Modification]
In the embodiment described above, the image pickup device 12 is arranged at a position for photographing the face of a passerby that passes through the automatic ticket gate 11. The position (photographing point) where the image pickup device 12 is arranged is not limited to the arrangement position of the automatic ticket gate 11. For example, a position where the face of a passerby that passes through a gate device arranged at an entrance or exit of a facility other than a station is photographed may be selected.

  Further, the shooting point is not limited to the arrangement position of the gate device, and for example, a park or a street corner may be selected as the shooting point. When the image pickup device 12 is arranged at a photographing point where a gate device such as a park or a street corner is not arranged, a passage detection sensor that detects passage of a passerby is arranged at the photographing point, and a passerby passes by the passage detection sensor. The image pickup device 12 is configured to take a picture of a passerby by using the detection of the above as a trigger. In addition, in this modification, since the passerby does not pass using the ticket, in the processing by the grouping unit 133 (1) When the ticket ID is not available and the entrance is passed (FIG. 8, S101- Only processing similar to S105) is performed. Further, the processing by the pairing unit 134 is not performed.

  The data structure, processing order, quantity, and the like described in the above embodiments and modifications are merely schematically illustrated to the extent that the present invention can be understood and implemented. Therefore, the present invention is not limited to the described embodiments, and can be variously modified without departing from the scope of the technical idea shown in the claims.

DESCRIPTION OF SYMBOLS 1 ... Movement path | route identification system, 9 ... Network, 11 ... Automatic ticket gate, 12 ... Image pick-up machine, 13 ... Movement path | route identification apparatus, 14 ... Memory | storage device, 15 ... Terminal device, 111 ... Pass sensor, 112 ... Transmission part, 113 DESCRIPTION OF SYMBOLS ... Reception part 114 ... Timekeeping part 115 ... Generation part 116 ... Transmission part 117 ... Reading part 121 ... Reception part 122 ... Imaging part 123 ... Transmission part 131 ... Reception part 132 ... Facial feature specific part DESCRIPTION OF SYMBOLS 133 ... Grouping part, 134 ... Pairing part, 135 ... Movement path | route identification part, 136 ... Transmission part, 151 ... Operation part, 152 ... Transmission part, 153 ... Reception part, 154 ... Display part

Claims (7)

Photographing point passage data indicating at least the characteristics of the face of a person photographed at each of a plurality of photographing points, the time of the photographing, and the location of the photographing, and the face characteristics indicated by the photographing point passage data are the same person Grouping means for grouping each of the items indicating, and storing in a storage device;
An apparatus comprising: a specifying unit configured to specify a movement path of a face person having a feature indicated by shooting point passage data based on a plurality of grouped shooting point passage data items. Each of the plurality of shooting points is an entrance / exit, and the shooting point passage data indicates whether entry or exit,
Pairing means for pairing shooting point passage data indicating entry and participation as a pair based on the time indicated by the shooting point passage data and the entry and exit among the shooting point passage data grouped into the same group Prepared,
The apparatus according to claim 1, wherein the specifying unit specifies the movement path of the person based on paired shooting point passage data. Each of the plurality of shooting points is a ticket gate of a station,
The grouping means, when grouping the shooting point passage data indicating the participation, based on the time and place indicated by the shooting point passage data and the time indicating the time required to move between the stations. The apparatus according to claim 2, wherein a group candidate for grouping is specified. The shooting point passing data indicates identification information of the ticket used by the person to pass the shooting point,
When grouping the shooting point passage data, the grouping unit, as a grouping destination group candidate of the shooting point passage data, indicates a shooting point indicating the same ticket identification information as the ticket identification information indicated by the shooting point passage data. The apparatus according to any one of claims 1 to 3, wherein a group including passage data is specified. A part of the shooting point passage data indicates the identification information of the ticket used by the person to pass the shooting point,
The grouping means, when grouping the shooting point passage data, if the shooting point passage data indicates ticket identification information, is the same as the identification information of the ticket as a grouping destination group of the shooting point passage data. After specifying a group including shooting point passage data indicating ticket identification information, grouping based on similarity of facial features is performed using a lower threshold than when shooting point passage data does not indicate ticket identification information. Item 5. The apparatus according to any one of Items 1 to 4. The grouping unit generates a new group including the photographing point passage data when there is no grouping destination group of the photographing point passage data when grouping the photographing point passage data. The device according to item. The grouping means, when the elapsed time from the latest time among the generated plurality of groups indicated by the shooting point passing data included in the group satisfies a predetermined condition, stores the data of the group in the storage device. The apparatus according to claim 1, wherein the apparatus is deleted.

Images