JP4569190B2 - Suspicious person countermeasure system and suspicious person detection device - Google Patents

Description

  The present invention relates to a technique that is effective when applied to a system or a method for maintaining the security of an object or a target area by capturing an image in a monitoring range and analyzing the image.

  In recent years, interest in security has increased, and for example, improvement in security in cars has been attracting attention. From the viewpoint of security in vehicles, there is a demand for technology for crime prevention, particularly against vandalism and mischief on the body.

Conventionally, as a crime prevention technique as described above, a technique has been proposed in which the outside of a vehicle is periodically imaged and the approach of a moving object to the vehicle is determined based on the difference between images captured before and after ( Patent Document 1). In addition, a technique has been proposed in which an approach of a human body is detected by detecting a change in a physical quantity around the vehicle, and threatening notification is executed so as to prevent theft on the vehicle (see Patent Document 2). Furthermore, a technique has also been proposed in which when an unauthorized operation of a vehicle by a third party is detected, a message to that effect is transmitted to an information center together with vehicle information representing the current vehicle state (see Patent Document 3).
JP-A-11-334536 JP 2003-99863 A JP 2003-157383 A

  However, with these conventional technologies, is the person approaching the vehicle a third party to be wary of, or whether it is a party that does not need to be wary (the owner of the vehicle, its family, friends, etc.)? It is not judged at all. For this reason, even if a person who does not need to be alerted approaches the vehicle, there is a risk that this approach will be detected and unnecessary alerts and threats will be implemented. In such a case, a legitimate operation by a related person is hindered. In addition, in the conventional technology, there is a possibility that even a non-malicious third party crossing the vicinity of the vehicle is detected and an unnecessary threat is executed. In such a case, it may not only give an unpleasant impression to a non-malicious third party, but may also give unnecessary effort to the owner who senses this threat. As a result, the reliability of the sensor itself is lowered.

  Therefore, the present invention solves these problems, and does not merely detect the approach of a person or the like and perform a warning, but accurately determines the situation to be warned and performs a process such as warning or intimidation as necessary. An object of the present invention is to provide a device.

  First, as one example of a situation to be wary of, there is a situation where a person who is very close to the suspicious person countermeasure system is detected. In this case, there is a high possibility that the imaging means in the suspicious person countermeasure system images the face of the person very large. Therefore, when a person's face is imaged very large in this way, the above problem can be solved by executing a countermeasure process for a suspicious person as a situation to be wary.

Based on such a viewpoint, the first aspect of the present invention is a suspicious person countermeasure system, which includes an imaging means, a detection means, a determination means, and a countermeasure processing means. The imaging means images a scene in the monitoring range. The monitoring range varies depending on the situation where the suspicious person countermeasure system is installed. For example, when a suspicious person countermeasure system is installed in a vehicle, the monitoring range may be set as the periphery of the vehicle. For example, when a suspicious person countermeasure system is installed in a house,
The monitoring range may be set around the house, on the garden or on the roof. The detection unit detects a human face from the image captured by the imaging unit, and acquires the size of the face. The detection means does not necessarily need to acquire the value of the face size itself as the face size, and may acquire another value closely related to the face size as the face size. Other values closely related to the size of the face are values such as the distance between the left and right eyes of the subject, the distance between the nostrils, and the size of the eyes or pupils. Further, the size of the face acquired by the detecting means is not the actual size of the face but the size on the image to be processed. In other words, the size of the face acquired by the detection means is a value that changes according to the position of the face in the real world. The determining means determines whether or not countermeasure processing for the suspicious person is necessary based on the size of the face of the person detected by the detecting means. For example, the determination unit may determine that countermeasure processing for a suspicious person is necessary when the size of a person's face is larger than a preset threshold. The countermeasure process includes a process for keeping the suspicious person away and a process for notifying the presence of the suspicious person to surrounding persons, users, security company persons, and the like. The countermeasure processing means performs countermeasure processing according to the determination by the determining means.

  In the first aspect of the present invention configured as described above, whether or not countermeasure processing is necessary is determined based on the size of the face detected by the detection means. For this reason, it is possible to prevent the countermeasure processing from being performed unnecessarily when the countermeasure processing is unnecessary even if a face is detected, for example, when a distant face is detected. It becomes. Further, for example, when the imaging means is configured to take an image through the window, the size of the face imaged to such an extent that the person of the face detected by the detection means can be recognized as looking into the inside of the window. If larger, the determination unit may be configured to determine that countermeasure processing is necessary. With this configuration, the first aspect of the present invention can implement the countermeasure process only when it is recognized that the detected face person is looking into the window.

  In the first aspect of the present invention, the situation to be warned is determined based on the size of the face at that time, but it seems that a person staying in the monitoring range of the suspicious person countermeasure system for a long time is detected. It is also possible to determine a situation to be warned from information accumulated for each detected face person, such as a situation where there is a person who has been detected many times in the monitoring range.

  Based on such a viewpoint, the second aspect of the present invention is a suspicious person countermeasure system, which includes an imaging means, a detection means, a feature amount storage means, an identification means, a storage means, a judgment means, and a countermeasure processing means. . The imaging unit images a scene in the monitoring range, similar to the imaging unit in the first aspect of the present invention. The detection unit detects a human face from the image captured by the imaging unit. The feature quantity storage means stores the feature quantity of each face detected by the detection means. The identification unit recognizes each face individually by performing identification processing on each face detected by the detection unit using the feature amount stored in the feature amount storage unit. The storage means accumulates and stores information related to each face identified by the identification means. The determination means determines whether or not countermeasure processing for the suspicious person is necessary based on the information stored in the storage means. The countermeasure processing means implements the countermeasure processing in accordance with the determination by the determination means, like the countermeasure processing means in the first aspect of the present invention.

  In the second aspect of the present invention configured as described above, the necessity of the countermeasure processing is determined based on the information regarding each face accumulated and stored in the storage unit. Therefore, it is possible to more accurately determine the necessity of the countermeasure process than when the necessity of the countermeasure process is determined based only on the detection result at that time. For example, it is possible to prevent the countermeasure processing from being performed unnecessarily for a person who happens to step into the monitoring range or a person who passes by.

Further, the storage means in the second aspect of the present invention may be configured to store, as information, the time when each face is continuously detected and / or the number of times each face appears in the monitoring range. . In this case, the determination unit may be configured to determine whether countermeasure processing is necessary based on the time and / or the number of times. Note that the number of times each face appears in the monitoring range does not indicate the number of times the face is detected by the detection means, but a face that has not been detected in the monitoring range for a predetermined time (detected by the first detection). This is the number of times counted as one time when a new or anew is detected within the monitoring range.

  In the second aspect of the present invention configured as described above, the necessity of the countermeasure processing is determined based on the time when each face is continuously detected, the number of times each face appears in the monitoring range, and the like. . For this reason, a person who is recognized as a suspicious person, such as a person who stays in the monitoring range for a long time, or a person who stays in the monitoring range for a short time, but who has repeatedly entered and exited the monitoring range. It is possible to execute countermeasure processing only for this. Therefore, unnecessary countermeasure processing can be prevented.

  Further, the detection means in the second aspect of the present invention may be configured to further acquire the size of the detected face of the person. The detection means does not necessarily need to acquire the value of the face size itself as the face size, and may acquire another value closely related to the face size as the face size. Further, the size of the face acquired by the detecting means is not the actual size of the face but the size on the image to be processed. That is, the size of the face acquired by the detecting means changes according to the position of the face in the real world. In this case, the determination means may be configured to determine whether countermeasure processing is necessary based on the size of the face. In other words, the determination means may be configured to determine whether countermeasure processing is necessary based on the size of the face and the time and / or the number of times.

  In the second aspect of the present invention configured as described above, the size of the face in the captured image is also one of the criteria for determining whether countermeasure processing is necessary. For this reason, it is possible to more accurately determine whether countermeasure processing is necessary. For example, it is possible to distinguish between a person who stays at a position far from the imaging apparatus for a long time and a person who stays at a position near the imaging apparatus for a long time, and determines whether countermeasure processing is necessary.

  Further, the feature amount storage means in the second aspect of the present invention further stores a face feature amount registered in advance, and stores the attribute information of the person of the face in association with the feature amount of each face. It may be configured to. The attribute information may indicate, for example, whether the person is a safe person (such as a person who does not need to be wary) or a dangerous person (a person who should be wary). When the feature amount storage unit is configured in this way, the identification unit may be configured to acquire the attribute information of the person of the face by performing an identification process on each face detected by the detection unit. . And when the 2nd aspect of this invention is comprised in this way, a judgment means may be comprised so that the necessity of countermeasure processing may be judged further based on attribute information.

  In the second aspect of the present invention configured as described above, the attribute information of the captured person is also one of the criteria for determining whether countermeasure processing is necessary. For this reason, it is possible to more accurately determine whether countermeasure processing is necessary. For example, it is possible to prevent unnecessary countermeasure processing from being performed on a safe person who does not need to perform countermeasure processing.

  Further, the determination means in the first or second aspect of the present invention may be configured to further determine the content of the countermeasure process to be executed when the countermeasure process is necessary. When configured in this way, the countermeasure processing means may be configured to execute the countermeasure processing according to the content of the countermeasure processing determined by the determining means. By being configured in this way, it is possible to prevent the countermeasure processing with the contents depending on the situation from being executed and executing the excessive countermeasure processing.

According to a third aspect of the present invention, there is provided a suspicious person detection device that detects a person's face from an image obtained by capturing a scene in the monitoring range, and detects the size of the face. An instruction means for determining whether or not countermeasure processing for the suspicious person is necessary based on the size of the face of the person who has been sent, and instructing the countermeasure processing device for executing the countermeasure processing to execute the countermeasure processing according to the determination including. By configuring the suspicious person countermeasure system shown as the first aspect of the present invention using the third aspect of the present invention configured as described above, the same effect as the first aspect of the present invention is obtained. It becomes possible.

  According to a fourth aspect of the present invention, there is provided a suspicious person detection device that detects a human face from an image obtained by capturing a scene in a monitoring range, and stores a feature amount of each face detected by the detection means. The feature amount storage means, the identification means for each face detected by the detection means using the feature amount, the identification means for individually recognizing each face, and the information about each face identified by the identification means Based on the result of the determination, the storage means for storing and storing, and determining whether or not the countermeasure processing for the suspicious person is necessary based on the information stored in the storage means, Instruction means for instructing execution of processing. By configuring the suspicious person countermeasure system shown as the second aspect of the present invention using the fourth aspect of the present invention configured as described above, the same effects as those of the second aspect of the present invention are obtained. It becomes possible.

  The third or fourth aspect of the present invention may be realized by executing a program by an information processing apparatus. That is, the present invention specifies the processing executed by each means in the third or fourth aspect of the present invention as a program for causing the information processing apparatus to execute or a recording medium on which the program is recorded. be able to. In addition, the present invention may be specified by a method in which the information processing apparatus executes the processing executed by each unit according to the third or fourth aspect of the present invention.

  In the present invention, the necessity of countermeasure processing is determined based on the size of the face detected by the detection means, information accumulated about the person of the face detected by the detection means, and the countermeasure processing is performed according to the determination result. Executed. For this reason, it is possible to prevent the countermeasure processing from being performed unnecessarily.

[System configuration]
The suspicious person countermeasure system 1 detects a suspicious person who has entered the monitoring range, and performs countermeasure processing for crime prevention in accordance with the degree of danger estimated from information related to the suspicious person. For example, when the suspicious person countermeasure system 1 is provided in a vehicle, the suspicious person approaching the vehicle is detected with the periphery of the vehicle as a monitoring range. Further, for example, when the suspicious person countermeasure system 1 is provided in a house, the suspicious person who has entered the roof or the garden is detected by using the yard or roof of the house as a monitoring range.

  First, the system configuration of the suspicious person countermeasure system 1 will be described. FIG. 1 is a diagram illustrating a functional block example of the suspicious person countermeasure system 1. The suspicious person countermeasure system 1 includes an imaging device 2, a suspicious person detection device 3, and a countermeasure processing device 4. Hereinafter, each device included in the suspicious person countermeasure system 1 will be described.

[Imaging device]
The imaging device 2 captures a scene in a range monitored by the suspicious person countermeasure system 1 and inputs data of the image to the suspicious person detection device 3. Any existing technology may be applied to the imaging device 2 as long as it can capture an image and input the image to the suspicious person detection device 3. For example, the imaging device 2 may be configured by a digital still camera, a digital video camera, or the like using a CCD (Charge-Coupled Devices), a CMOS (Complementary Metal-Oxide Semiconductor) sensor, or the like.

  The imaging device 2 may be configured to perform imaging at a timing set by a designer or a user. For example, when the imaging apparatus 2 is configured as a digital still camera, the imaging apparatus 2 may be configured to perform imaging at a predetermined time interval (every second, every minute, etc.). For example, when the imaging device 2 is configured as a digital video camera, the imaging device 2 may be configured to continuously perform imaging and input image data of a frame at a predetermined timing to the suspicious person detection device 3. Further, the imaging device 2 may be configured to capture scenes in different ranges (areas) at predetermined time intervals. Further, only one imaging device 2 may be provided in the suspicious person countermeasure system 1, or a plurality of imaging devices 2 may be provided. When a plurality of imaging devices 2 are provided, each imaging device 2 may be configured such that data input to the suspicious person detection device 3 does not overlap by performing imaging at different timings.

[Suspicious person detection device]
The suspicious person detection device 3 performs processing using image data (hereinafter referred to as “input image”) input from the imaging device 2, thereby detecting a person existing within the monitoring range of the suspicious person countermeasure system 1. To detect. When the suspicious person detection device 3 detects a person, the suspicious person detection device 3 estimates the degree of danger (warning level) for the person, and instructs the countermeasure processing device 4 to execute countermeasure processing as necessary. The suspicious person detection device 3 estimates the degree of danger for each detected person according to the frequency of detection, the time continuously detected, the personal attribute of the person, and the like.

The suspicious person detection device 3 includes a CPU (Central Processing Unit), a main storage device (RAM), an auxiliary storage device, and the like connected via a bus in terms of hardware. The auxiliary storage device is configured using a nonvolatile storage device. The nonvolatile storage device referred to here is a so-called ROM (Read-Only Memory: including EEPROM (Erasable Programmable Read-Only Memory), EEPROM (Electrically Erasable Programmable Read-Only Memory), mask ROM, etc.), FRAM (Ferroelectric RAM). ), Hard disk, etc. The suspicious person detection device 3 loads various programs (OS, applications, etc.) stored in the auxiliary storage device into the main storage device and is executed by the CPU, so that the face detection unit 5, the database storage unit 6, person tracking It functions as a device including the unit 7, the face collation unit 8, the warning level determination unit 9, and the like. The face detection unit 5, the person tracking unit 7, the face collation unit 8, and the alert level determination unit 9 may be realized by executing a program by the CPU. Further, the face detection unit 5, the person tracking unit 7, the face matching unit 8, and the warning level determination unit 9 may be configured as a dedicated chip. Next, each functional unit included in the suspicious person detection device 3 will be described.

<Face detection unit>
The face detection unit 5 detects a human face from the input image and specifies face information indicating the position, size, orientation, and the like of the detected face. For example, the face detection unit 5 may be configured to detect a face by template matching using a reference template corresponding to the outline of the entire face. Further, the face detection unit 5 may be configured to detect a face by template matching based on face components (eyes, nose, ears, etc.). The face detection unit 5 may be configured to detect a vertex such as a head by chroma key processing and detect a face based on the vertex. The face detection unit 5 may be configured to detect a region close to the skin color and detect the region as a face. Further, the face detection unit 5 may be configured to perform learning by a teacher signal using a neural network and detect a face-like area as a face. Further, the face detection process by the face detection unit 5 may be realized by applying any other existing technique. The skin color need not be limited to the skin color of a specific race, and may be configured to be compatible with the skin color of any race. For example, the setting may be made according to the region where the suspicious person countermeasure system 1 is used, or the skin color may be specified from the input image.

  The face detection unit 5 passes face information about each detected face to the person tracking unit 7.

<Database storage unit>
The database storage unit 6 stores a personal database and a temporary database. FIG. 2 is a diagram illustrating examples of the personal database and the temporary database. FIG. 2A shows an example of a personal database, and FIG. 2B shows an example of a temporary database.

  The personal database has information about each individual as a record. Specifically, the personal database has ID, attribute, frequency, and face information as items. The ID is an identifier assigned to each individual, and is data for associating the personal database with the temporary database. The attribute is data indicating whether or not the person should be wary, and has values such as “safe” and “danger”, for example. The frequency is the number of times when a face that has not been detected within the monitoring range for a predetermined time or longer (except when it is detected by the first detection) is newly or newly detected within the monitoring range is counted as one time. Show. For example, the frequency value is updated (for example, incremented) when a person who has been detected continuously for a predetermined time is no longer detected and then continues to be detected again for a predetermined time. The face information is data used by the face matching unit 8 for face identification processing, and may be, for example, a face feature amount acquired from a face image or an actual face image. Note that the frequency value need not be held for a person whose attribute is “safe”.

  The temporary database is created for each individual to be tracked by the person tracking unit 7 and accumulates detection and tracking results. Each temporary database has a corresponding individual ID. Specifically, the temporary database has face images, sizes, and times as items. The face image is a face image of a person to be tracked, that is, an individual face image corresponding to each temporary database, and has face image data detected by the face detection unit 5. Although all the detected face images may be accumulated, for the purpose of saving storage capacity, a configuration may be adopted in which several of the latest face images are stored. The size indicates the size of the face in each face image. The face size is a value included in the face information acquired by the face detector 5. The time indicates a time when an input image including each face image is captured, a time when the image is input to the suspicious person detection device 3, and the like.

<Person Tracking Section>
The person tracking unit 7 requests the face matching unit 8 to execute a matching process (identification process) for each face person detected by the face detecting unit 5. The person tracking unit 7 updates the temporary database based on the result of the collation process, and manages the contents of the temporary database of the person to be tracked. Then, the person tracking unit 7 requests the warning level determination unit 9 to acquire a warning level for an arbitrary person as necessary. Specific processing performed by the person tracking unit 7 will be described in the column of an operation example.

<Face matching unit>
The face matching unit 8 performs an identification process on the person of the face image requested by the person tracking unit 7. When performing the identification process, the face matching unit 8 uses the face information included in the personal database to determine whether or not the person in the face image is included in the personal database, and if so, which person It is judged whether it is. For example, the face matching unit 8 can perform the identification process by using a technique described in the following document.

Takio Kurita, “Statistical Method for Face Detection / Face Recognition”, [Search June 4, 2004], Internet <URL: http://www.neurosci.aist.go.jp/~kurita/lecture /statface.pdf
>
Kazuhiro Fukui (Toshiba Corporation), “Face Image Recognition in the User Interface”, SSII2000 Tutorial Lectures, Page 18-32

<Warning level judgment part>
The warning level determination unit 9 determines the current warning level of the person requested by the person tracking unit 7 according to the contents of the personal database or the temporary database. The alert level is information indicating whether or not the countermeasure processing by the countermeasure processing device 4 is necessary or how much countermeasure processing is necessary in the range of the range to be monitored by the suspicious person countermeasure system 1. For example, several levels such as level A (no countermeasure processing is required), level B (simple countermeasure processing is required), level C (sufficient countermeasure processing is required) may be set as the alert level. FIG. 3 is a diagram illustrating an example of a warning level map used when the warning level determination unit 9 determines a warning level. The alert level map has stagnation time and frequency as axes, and is configured such that one alert level is determined according to the set of these two values.

  A plurality of alert level maps may be prepared according to the size of the detected face, for example. Specifically, the alert level determination unit 9 uses the alert level map of FIG. 3A when the size of the detected face is larger than the threshold, and when smaller than the threshold, FIG. It may be configured to use the alert level map of b). In FIG. 3A, the warning level is determined as level B or level C with a shorter stagnation time or less frequency than in FIG.

  The warning level determination unit 9 acquires three values of stagnation time, frequency, and face size when determining the warning level. The method of acquiring each value will be described in the operation example column. After acquiring these three values, the warning level determination unit 9 determines a warning level map used for processing based on the size of the face. And the alert level judgment part 9 determines an alert level based on the area | region where the point corresponding to the value of stagnation time and frequency is located in the determined alert level map.

  The alert level determination unit 9 passes the alert level as a determination result to the countermeasure processing device 4. At this time, the alert level determination unit 9 may be configured to pass the alert level to the countermeasure processing device 4 only when the alert level is level B or level C.

[Countermeasure equipment]
The countermeasure processing device 4 executes countermeasure processing according to the value of the alert level output from the suspicious person detection device 3. The countermeasure processing includes threatening processing for a suspicious person, warning processing for a user, notification via a network, and the like. For example, the countermeasure processing device 4 does not execute the process when the alert level is level A, and warns the suspicious person not to approach by suspicious person (for example, voice output or light emission) when the alert level is level B. If the alert level is C, a report via the network or public network (such as a report to the police or a report to the security service) or a warning to the user (user's mobile phone) Or notification to the mobile terminal by e-mail, notification by siren or light, etc.).

[Operation example]
4 to 6 are flowcharts showing an operation example of the suspicious person countermeasure system 1. An operation example of the suspicious person countermeasure system 1 will be described with reference to FIGS.

When the process is started, the imaging device 2 captures a scene in the monitoring range and acquires an image (S1). The face detection unit 5 detects a human face from the image captured by the imaging device 2 (S2). The face detection unit 5 passes the detection result, that is, the face information for each detected face to the person tracking unit 7. When the face tracking unit 7 receives face information from the face detection unit 5, that is, when a face is detected by the face detection unit 5 (S <b> 3 -YES), the person tracking unit 7 performs matching on the face person to the face matching unit 8. Upon request, the face matching unit 8 performs face matching based on the request (S4). The face collation unit 8 passes the collation result to the person tracking unit 7. On the other hand, when the face tracking unit 7 does not receive face information from the face detection unit 5, that is, when no face is detected by the face detection unit 5 (S3-NO), the processes after S1 are executed again.

  The person tracking unit 7 determines subsequent processing in accordance with the matching result received from the face matching unit 8. If the face person detected as a result of the collation is a person already registered in the temporary database (described as “temporary DB” in the flowchart) (S5-YES), the person tracking unit 7 The temporary database is updated (S6). Specifically, the person tracking unit 7 registers the captured face image, the size of the face in the face image, and the time when the face image was captured as a new record in the temporary database of the corresponding person. . Next, the person tracking unit 7 calculates the stagnation time of the person to be processed, and determines whether or not the time is equal to or greater than a threshold value (S7). The person tracking unit 7 acquires the difference between the time in the oldest record in the temporary database and the time in the newest record as the stagnation time. When the stagnation time is equal to or greater than the threshold (S7-YES), the person tracking unit 7 requests the warning level determination unit 9 to acquire the warning level for the person corresponding to the temporary database. Then, the alert level determination unit 9 executes the processes after S9 based on the records of the temporary database and personal database (described as “personal DB” in the flowchart) corresponding to the requested person. The processing after S9 will be described later.

  In the process of S5, when the person of the face image to be processed is not registered in the temporary database (S5-NO), the person tracking unit 7 creates a temporary database for this person (S8). Then, the processes after 016 are executed. Moreover, in the process of S7, when the stagnation time is less than a threshold value (S7-NO), the process after S16 is performed. The processing after S16 will be described later.

  The processing after S9 will be described. First, the alert level determination unit 9 acquires the stagnation time for the person who is the object of processing (S9). At this time, the alert level determination unit 9 may acquire the stagnation time already calculated from the person tracking unit 7 (that is, the stagnation time calculated in the process of S7), or may be calculated again from the temporary database. Also good. Moreover, the warning level determination part 9 acquires the frequency about the person who is the object of processing (S10). The alert level determination unit 9 acquires the frequency of the person who is the object of processing from the personal database. Moreover, the alert level determination part 9 acquires the face size about the person who is the object of processing (S11). The alert level determination unit 9 acquires the face size from the latest record in the temporary database for this person. And the alert level judgment part 9 determines the present alert level about the person who is the object of a process based on these acquired three values (S12).

  If the determined alert level is not level A (S13-NO), the alert level determination unit 9 acquires the attribute of the person who is the object of processing from the personal database. When the attribute of this person is dangerous (S14-danger), the warning level determination unit 9 outputs the warning level to the countermeasure processing device 4. And the countermeasure processing apparatus 4 implements a countermeasure process according to the alert level output from the alert level judgment part 9 (S15). If the alert level is level A in the process of S13 (S13-YES), if the attribute of the person is safe in the process of S14 (S14-safe), after the process of S15, after the process of S8, Or in S7-NO, the process of S16 is performed. In other words, the person tracking unit 7 determines whether the processes in S4 to S15 have been completed for all the faces detected by the face detecting unit 5 (S16). When there is a face that has not been processed yet (S16-NO), the process after S4 is performed on the face that has not been processed.

On the other hand, when processing has been completed for all detected faces (S16-YES), the person tracking unit 7 determines whether there is a temporary database that is not referenced in the processing of S4 to S15 (S17). If there is such a temporary database (S17-YES), the person tracking unit 7 determines whether or not a predetermined time or more has elapsed since the last update for each temporary database that is not referenced (S18). That is, the person tracking unit 7 determines whether or not a predetermined time or more has elapsed since the time of the latest record in each temporary database. When the predetermined time or more has elapsed (S18-YES), the person tracking unit 7 updates the record by incrementing the frequency value of the record corresponding to the temporary database among the records in the personal database. Perform (S19). The person tracking unit 7 performs the processes of S18 and S19 for all corresponding temporary databases. If there is no unreferenced temporary database in S17 (S17-NO), or if there is no non-referenced temporary database after a predetermined time has elapsed since the last update (S18-NO), or S19 After the process, the processes after S1 are executed.

[Action / Effect]
In the suspicious person countermeasure system 1, a suspicious person is detected by detecting a person's face from an image obtained by imaging a monitoring area. Next, identification processing is performed on the detected face, and the face of a safe person (for example, a relative or friend) or a dangerous person (for example, an unknown person or a person registered as a dangerous person) Whether it is a face is determined. Only when it is determined that the face is a dangerous person, the countermeasure processing device 4 performs the countermeasure process. For this reason, countermeasure processing is not performed for all persons who have entered the monitoring range, but countermeasure processing is performed only for those who are determined to be dangerous persons. Therefore, it is possible to prevent unnecessary countermeasure processing from being performed on a safe person.

  Further, in the suspicious person countermeasure system 1, the alert level is determined according to the time during which the suspicious person is continuously detected (stagnation time) and the frequency at which the suspicious person is detected. And the countermeasure processing apparatus 4 changes the grade of the countermeasure process to perform according to the determined alert level. For this reason, since the content of the countermeasure process changes according to the situation, it is possible to prevent the unnecessary countermeasure process from being performed unnecessarily.

[Modification]
The alert level determination unit 9 is configured to determine the alert level as level B or level C regardless of the alert level map when the size of the human face detected by the face detection unit 5 is equal to or greater than the threshold value. May be. With this configuration, it is possible to more flexibly implement countermeasure processing for a suspicious person. For example, even for a suspicious person suddenly appearing in front of the imaging apparatus 2, that is, a suspicious person with very small stagnation time and frequency values, the suspicious person countermeasure system 1 performs countermeasure processing according to the size of the face. It becomes possible to carry out.

  Further, the alert level map does not necessarily have to be two-dimensional or three-dimensional, and may be configured in one dimension (for example, only one of stagnation time or frequency), or configured as a three-dimensional map or more. Also good.

  The person tracking unit 7 may be configured not to detect a person's face but to track a person based on other information about the person. For example, tracking of a person may be realized based on information about clothes. Such tracking is, for example, a technique (“http://www.ced.co.jp”) that allows object movement information in a three-dimensional space to be extracted from velocity field information and accurately identifies a moving object. The technology described in “/riv3.htm” etc.).

  Further, the processing of S14 in FIG. 5 (determination processing for determining the subsequent processing in accordance with the attribute of the person to be processed) is performed immediately after the processing of S4 (face matching processing) in FIG. It may be performed at the timing.

It is a figure which shows the example of a functional block of a suspicious person countermeasure system. It is a figure which shows the example of each of a personal database and a temporary database. It is a figure which shows the example of a vigilance level map. It is a flowchart which shows the operation example of a suspicious person countermeasure system. It is a flowchart which shows the operation example of a suspicious person countermeasure system. It is a flowchart which shows the operation example of a suspicious person countermeasure system.

Explanation of symbols

DESCRIPTION OF SYMBOLS 1 Suspicious person countermeasure system 2 Imaging device 3 Suspicious person detection apparatus 4 Countermeasure processing apparatus 5 Face detection part 6 Database storage part 7 Person tracking part 8 Face collation part 9 Warning level judgment part

Claims (5)

Imaging means for imaging a scene in the monitoring range;
Detecting means for detecting a human face from an image captured by the imaging means;
Feature quantity storage means for storing the feature quantity of each face detected by the detection means;
Identifying means for individually recognizing each face by performing identification processing using the feature amount for each face detected by the detecting means;
Storage means for storing and storing information about each face identified by the identification means;
Determination means for determining whether or not countermeasure processing for the suspicious person is necessary based on the information stored in the storage means;
Countermeasure processing means for performing the countermeasure processing according to the determination by the determination means,
The detection means further acquires the size of the detected human face on the image of the face,
It said determination means has a plurality of alert level map as a reference for determining necessity of countermeasure processing depending on whether the size of the image of the face is greater than a predetermined threshold, countermeasure processing The suspicious person measures system characterized by changing the warning level map used for judgment of necessity . The storage means stores, as the information, a time when each face is continuously detected as a stagnation time, and stores the number of times each face appears within the monitoring range as a frequency ,
The determining means is a suspicious person measures system of claim 1, wherein the dwell time based on及beauty alert level map that is configured in the previous SL frequency, to determine the necessity of the countermeasure process. Detection means for detecting a human face from an image obtained by capturing a scene in the monitoring range;
Feature quantity storage means for storing the feature quantity of each face detected by the detection means;
Identifying means for individually recognizing each face by performing identification processing using the feature amount for each face detected by the detecting means;
Storage means for storing and storing information about each face identified by the identification means;
Based on the information stored in the storage means, the necessity of countermeasure processing for the suspicious person is determined, and the countermeasure processing device that executes the countermeasure processing is instructed to execute the countermeasure processing according to the result of the determination An instruction means for
The detection means further acquires the size of the detected human face on the image of the face,
It said indicating means includes a plurality of alert level map as a reference for determining necessity of countermeasure processing depending on whether the size of the image of the face is greater than a predetermined threshold, countermeasure processing A suspicious person detecting device, wherein the warning level map used for determining whether or not the information is necessary is changed. A detection step of detecting a human face from an image obtained by capturing a scene of the monitoring range;
Storing the detected feature quantities of each face;
Recognizing each face individually by performing an identification process using the feature amount for each detected face;
Storing and storing information about each identified face;
An instruction step for determining whether or not countermeasure processing for the suspicious person is necessary based on the stored information, and instructing the countermeasure processing device that executes the countermeasure processing to execute the countermeasure processing according to the result of the determination; Is a program for causing the information processing apparatus to execute
In the detection step, for the detected human face, the size of the face on the image is further acquired,
Wherein in the instruction step, it has a plurality of alert level map as a reference for determining necessity of countermeasure processing depending on whether the size of the image of the face is greater than a predetermined threshold, countermeasure processing A program characterized by changing the alert level map used for determining whether or not it is necessary . A detection step in which the information processing device detects a human face from an image obtained by capturing a scene in the monitoring range;
An information processing device storing the detected feature amount of each face;
A step of individually recognizing each face by performing an identification process using the feature amount for each detected face by the information processing apparatus;
An information processing device storing and storing information about each identified face;
The information processing apparatus determines whether or not the countermeasure process for the suspicious person is necessary based on the stored information, and executes the countermeasure process according to the determination result to the countermeasure processing apparatus that executes the countermeasure process. An instruction step for indicating, and
In the detection step, for the detected human face, the size of the face on the image is further acquired,
Wherein in the instruction step, it has a plurality of alert level map as a reference for determining necessity of countermeasure processing depending on whether the size of the image of the face is greater than a predetermined threshold, countermeasure processing The suspicious person detection method characterized by changing the said alert level map used for the judgment of necessity of the .

Images