JP5375686B2 - Reporting device and reporting system - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a reporting apparatus that prevents complication in the duties of guards or others and prevent delay in dealing with persons who need to be dealt with quickly, by avoiding reporting persons who do not need to be dealt with quickly and reliably reporting persons who need to be dealt with quickly. <P>SOLUTION: An image processing unit 22 processes frame images captured by a camera 1 imaging a monitored area to analyze an imaged person's behavior and determine whether the behavior is equivalent to any of a plurality of predetermined detection acts. A habitual offender DB 24 stores the frequency of detection of each detection act about a person whose detection act has been asserted by the image processing unit 22. A threshold frequency is set for each detection act, and an output part 25 reports a person whose frequency of any detection act exceeds the corresponding threshold. <P>COPYRIGHT: (C)2011,JPO&amp;INPIT

Description

  The present invention relates to a notification device and a notification system for processing a frame image captured by an imaging device, analyzing the behavior of the person being imaged, and notifying a security guard or the like of a person who has caused a behavior related to fraud .

  Conventionally, the frame image captured by the camera is processed, the behavior of the captured person is analyzed, the person whose behavior is abnormal behavior including fraud etc. is detected, and the person is guarded To report to.

  It also captures and stores facial images of people who are likely to be cheating and uses face authentication technology to determine whether the person is an addict who is cheating. In addition, it is also proposed to notify a person who is an addict (see Patent Document 1).

JP 2007-143738 A

  However, every time a person whose behavior is abnormal behavior is detected, the configuration is such that the person is notified to a security guard or the like. As a result, the work of security guards and the like becomes complicated, and the number of manpower required for handling the notified person increases unnecessarily.

  Moreover, if it is set as the structure which reports about the person with an addictive property like patent document 1, the report of persons with a bad physical condition etc. which a guard needs to respond immediately is not performed. For this reason, the response | compatibility with respect to the person who must respond quickly may be overdue.

  The purpose of this invention is to refrain from reporting about persons who do not need to respond promptly, and to ensure that reports about persons who need to respond promptly are made, so that the work of security guards and the like becomes complicated. It is an object of the present invention to provide a notification device and a notification system capable of preventing and preventing delays in response to a person who must respond quickly.

  In order to solve the above-described problems and achieve the object, the reporting device of the present invention is configured as follows.

The behavior analysis means processes the frame image captured by the imaging device that captures the monitoring area, analyzes the behavior of the person being imaged, and the behavior of the person is one of a plurality of types of detection behavior that is predetermined. It is judged whether it corresponds to. The detection behavior referred to here includes not only behavior related to fraud but also behavior due to poor physical condition or trouble of the device being used.

In addition, the storage means classifies the plurality of detection actions into an addictive detection event having a set threshold number of times, and other specific detection events, and the action analysis means detects the addictive detection event. The number of times of detection is stored for each person who is determined to be an act related to. Then, passing report means, person behavioral analysis unit was determined to be according act on the detection action is a specific sense event, and the threshold number of times the detection frequency is set by the detection behavior in any addictive sensed events Report on persons who exceed.

Therefore, for detection behavior that does not need to be dealt with immediately, the threshold number is increased as an addictive detection event, and detection behavior that needs to be dealt with immediately (behavior due to poor physical condition or trouble with the device being used) By making a specific detection event, it is possible to refrain from reporting about a person who does not need to respond immediately, and to promptly report a person who needs to respond immediately. Thereby, it is possible to prevent work such as security guards from becoming complicated. In addition, it is possible to prevent a delay in the response to a person who must respond quickly.

Further, authentication of a person who is determined that the action corresponds to any of a plurality of predetermined detection actions may be performed by face authentication using the face image of the person or by another method. May be.

In addition, it is preferable that a report is made even for a person whose total number of times that the behavior analysis unit has determined that the detection behavior is related to the addictive detection event exceeds the total threshold number. In this way, it is possible to prevent delays in reporting a person who is performing detection actions related to various types of fraud.

  According to the present invention, it is possible to prevent work such as security guards from becoming complicated, and to prevent delays in dealing with a person who must respond quickly.

It is the schematic which shows the structure of a report system. It is a block diagram which shows the structure of the principal part of a notification apparatus. It is a figure which shows a report table. It is a flowchart which shows operation | movement of a notification apparatus.

  Hereinafter, a notification system according to an embodiment of the present invention will be described.

  FIG. 1 is a schematic diagram showing a configuration of a notification system according to an embodiment of the present invention. This notification system includes a plurality of cameras 1 and a notification device 2. Each camera 1 images the assigned monitoring area. The monitoring area is, for example, a station platform (home area), a ticket gate (ticket gate area), a ticket office (ticket vending machine area), and a fence area in which a fence is installed. The camera 1 is a video camera that outputs about 30 frames per second of a frame image obtained by imaging a monitoring area. Each camera 1 inputs a frame image obtained by imaging the monitoring area to the notification device 2.

  For each camera 1, the notification device 2 processes the input frame image, analyzes the action of the person being imaged, and determines whether the action is one of the predetermined detection actions. In addition, the reporting device 2 determines the necessity of reporting to a security guard or the like for a person who has determined that the behavior is a detection behavior, and performs reporting when it is determined that reporting is necessary.

  FIG. 2 is a block diagram illustrating a configuration of a main part of the notification device. The reporting device 2 includes a control unit 20, an image input unit 21, an image processing unit 22, a storage unit 23, an addict database (hereinafter referred to as addict DB 24), and an output unit 25. . The control unit 20 controls the operation of the reporting device 2 main body.

  The image input unit 21 and the image processing unit 22 are provided for each camera 1 to be connected. The image input unit 21 includes a buffer that temporarily stores a frame image input from the corresponding camera 1 (connected camera 1). The image processing unit 22 processes the frame image input from the camera 1 (the frame image stored in the above-described buffer), analyzes the action of the person being imaged, and detects a plurality of detections in which the action is predetermined. Determine whether it corresponds to one of the actions. This detection behavior includes not only behavior related to fraud but also behavior associated with the person's poor physical condition, behavior associated with the occurrence of a failure of the user device, and the like.

  Some detection actions are set according to the monitoring area of the camera 1, and some detection actions are set regardless of the monitoring area of the camera 1. For example, if the monitoring area is the home area, it is determined that the person is rushing on (detecting action) when the moving speed of the person is high. Further, if the monitoring area is a ticket gate area, it is determined that the ticket gate is illegally passed (detected behavior) when the posture of the person is bent or jumped. Further, if the monitoring area is a ticket gate area or a ticket vending machine area, it is determined that a device trouble (detection behavior) occurs if a person has stopped for a long time. Also, if the monitoring area is a fence area, it is determined that the fence has been overcome (detection action) when there is a jump in the posture of the person. Furthermore, a person who falls down or crouches for a long time without being limited to a monitoring area is determined to have poor physical condition (detection behavior).

  The posture of the person may be performed using a known method such as silhouette image analysis or pattern matching.

  The memory | storage part 23 has memorize | stored the notification table 23a (refer FIG. 3) which matched the threshold frequency | count for every detection action. FIG. 3 shows an example in which the last boarding is 5 times, the ticket gate is illegally passed 4 times, the equipment trouble and poor physical condition are 1 time, and the fence is passed 3 times. In the following description, a detection action with a threshold count of 1 is referred to as a specific detection event, and a detection action with a threshold count of multiple is referred to as an addictive detection event. Moreover, the memory | storage part 23 has memorize | stored the total threshold frequency (for example, 7 times) separately from this report table 23a.

  In the addict DB 24, for each person whose behavior is determined to be a detection action related to the addictiveness detection event in the image processing unit 22, the person's face image, the face part of the person (face outline, eye , Nose, mouth, and the like) and the number of detections for each addictive detection event are registered.

  The output unit 25 transmits a notification mail to a mobile terminal or the like possessed by a security guard or the like.

  The operation of this notification system will be described below.

  FIG. 4 is a flowchart showing the operation of the notification device. In the notification device 2, each image processing unit 22 processes the frame image input from the connected camera 1, and analyzes the action of the person being imaged (s1). When any one of the image processing units 22 detects a person whose action is determined to be a detection action (s2), the reporting device 2 detects whether the detection action is an action related to a specific detection event, or detects addictiveness. It is determined whether the action is an event (s3).

  When the reporting device 2 determines that the action is related to the specific detection event in s3, the output unit 25 transmits a reporting email to the portable terminal or the like possessed by a security guard or the like (s4). This notification mail includes the whereabouts of the person whose action is determined to be a specific detection event (the monitoring area of the camera 1 corresponding to the image processing unit 22 that made this determination), the captured image of the camera 1, and the like. It is. Therefore, a security guard or the like who possesses the mobile terminal that has received the notification mail can respond quickly.

  In this way, security officers need to immediately respond to poor physical condition and notifications related to equipment failures, etc., so it will be made surely and promptly. Will not be delayed.

  When the notification device 2 transmits the notification mail in s4, the notification device 2 returns to s1 and repeats the above-described processing.

  If the reporting device 2 determines that it is not a specific detection event in s3, that is, determines that it is an addictive detection event, it cuts out the face image of the person from the captured image of the camera 1 (s5), and extracts the feature amount of the facial part (S6). The reporting device 2 searches the addict DB 24 using the feature amount of the facial part extracted in s6, and determines whether or not the corresponding person has already been registered (s7). If the reporting device 2 determines that it is not registered in the addict DB 24 in s7 (is an unregistered person), it registers this person in the addict DB 24 (s8). In s8, the number of detections of this person's face image (the face image cut out in s5), the facial part feature amount (the facial part feature amount extracted in s6), and the addictiveness detection event determined this time are set to one. The addict record is registered in the addict DB 24.

  If the reporting device 2 determines that it is registered in the addict DB 24 in s7, the reporting device 2 counts up the number of times of detection of the addictiveness detection event determined this time for the corresponding addict record (s9). The reporting device 2 determines whether or not the number of detected addictive detection events determined this time is equal to or greater than the threshold number associated with the addictive detection event (s10).

  When the reporting device 2 determines that the number of times is equal to or greater than the threshold number in s10, the reporting device 2 transmits a notification mail in s4. In other words, the reporting device 2 associates the number of detected addictive detection events (for example, rush-in, overpass, etc.) with a person who repeatedly repeats (for example, an addictive detection event). A person who has exceeded the threshold number of times)

  As you can see, when you detect a rush-in ride, over a fence, etc., you will not be notified immediately, but you will be repeated many times, and you will be notified to the guards when the threshold is reached. .

  Moreover, even if it determines with not being more than a threshold frequency in s10, the alerting | reporting apparatus 2 determines whether the sum total of the frequency | count that the addictiveness detection event was detected is more than a total threshold frequency about the person (s11). If the reporting device 2 determines that the number of times is equal to or greater than the total threshold number in s11, the reporting device 2 transmits a notification mail in s4. That is, the reporting device 2 reports a security guard or the like about a person who repeats various types of addictive detection events many times even if the number of addictive detection events is less than the threshold number.

  When the reporting device 2 determines that the number of times is less than the total threshold number in s11, the reporting device 2 returns to s1 and repeats the above-described processing.

  As described above, in this reporting system, when it is determined that the detection action is related to a specific detection event related to poor physical condition or equipment trouble, a security guard is immediately notified. Therefore, it is possible to prevent delays in the response of security guards or the like to a person who must respond quickly.

  In addition, it is possible to refrain from reporting about a person who does not need to respond immediately. Therefore, it is possible to prevent work such as security guards from becoming complicated.

  In addition, for those who repeat certain types of addictive detection events many times, and those who repeat various types of addictive detection events many times, they will be notified to security guards, etc. The person who takes the action can be reported to the guards and caught.

  In addition, a camera is installed at the automatic ticket gate installed at the ticket gate of the station, and the automatic ticket gate detects illegal activities such as unauthorized use of commuter passes, passage of friends, unauthorized use of children's tickets, and passage of no bills. When it does, it is good also as a structure which notifies the notification apparatus 2 of the face image of the person, the kind of fraudulent act, etc. which were performed. In this case, illegal activities such as unauthorized use of commuter pass, passage with friends, unauthorized use of children's tickets, passage of no bills, etc. may be set as the above-mentioned addictive detection event.

  The image input unit 21 and the image processing unit 22 described above may be provided on the camera 1 side instead of the reporting device 2 side. In this case, if the camera 1 is configured to notify the notification device 2 of the facial image, the feature amount of the facial part, and the type of the detected behavior detected this time, of the person who has determined that the analyzed behavior is the detected behavior. Good.

DESCRIPTION OF SYMBOLS 1 ... Camera 2 ... Notification apparatus 20 ... Control part 21 ... Image input part 22 ... Image processing part 23 ... Memory | storage part 23a ... Notification table 24 ... Adjunct database (addict DB)
25 ... Output section

Claims (4)

Processing the frame image where the imaging device imaged being imaged surveillance area, whether analyzing the behavior of the person being imaged, which corresponds to one of the sensing action of a plurality of types of actions are predetermined for the person A behavior analysis means for judging
The plurality of detection behaviors are classified into an addictive detection event having a set threshold number of times and a specific detection event other than that, and the behavior analysis means is applied to the detection behavior that is the addictive detection event . for the person it is determined that an act, a storage means for storing the number of detections by the detection action,
The person determined that the action analysis means is an action related to the detection action that is the specific detection event, and the number of detection times of the detection action in any of the addictive detection events exceeds a set threshold number A reporting device comprising a reporting means for reporting a person who has been The said action analysis means performs the authentication of the person who judged that the action of the person corresponds to any of the predetermined types of the detected actions using the face image of the person. Notification device. The said reporting means also reports the person whose total number of times that the behavior analysis means has determined that the detection behavior related to the addictiveness detection event exceeds the total threshold number of times. Reporting device. For each monitoring area, a plurality of imaging devices that image the monitoring area;
For each of the image pickup apparatus, whether to process the frame image whose imaging device imaged analyzes the behavior of the person being imaged, which corresponds to one of the sensing action of a plurality of types of actions are predetermined for the person A behavior analysis means for judging
The plurality of detection behaviors are classified into an addictive detection event having a set threshold number of times and a specific detection event other than that, and the behavior analysis means is applied to the detection behavior that is the addictive detection event . for the person it is determined that an act, storage means for storing the number of detections by the detection action and,
The person determined that the action analysis means is an action related to the detection action that is the specific detection event, and the number of detection times of the detection action in any of the addictive detection events exceeds a set threshold number A reporting device comprising a reporting means for reporting a particular person, and a reporting system.

Images