JP5981015B1 - Shoplifting crime scene recording device - Google Patents

Abstract

A shoplifting crime scene recording apparatus capable of accurately discriminating whether or not a person is a shoplifter before a crime is committed and recording a crime scene caused by the shoplifter. The shoplifting crime scene recording device 1 is visually recognized from the outside so that a shooting area or a virtual shooting area can be recognized by a shoplifter. A hidden camera 20 for photographing a region, an image processing means 32 for processing an image of a person photographed by the hidden camera 20, a judging means 33 for judging whether or not the photographed person is a shoplifter, and a shoplifter And a recording device 40 for recording the crime scene by. The discriminating means 33 discriminates the person as a shoplifter when the photographed person looks up the alerting camera 10 below the alerting camera 10, and the crime scene where the recording device 40 is photographed by the hidden camera 20. Are recorded for a predetermined time. [Selection] Figure 1

Description

  The present invention relates to a shoplifting crime scene recording apparatus.

  In various places such as department stores, surveillance cameras are installed to monitor and shoot shoplifters. However, shoplifters have a habit of confirming the position of the surveillance camera and its shooting area, entering the blind spot of the surveillance camera, and committing crimes such as shoplifting. Therefore, since the presence and shooting area of a normal surveillance camera are easily recognized by shoplifters and the like, it is difficult to photograph the crime scene. For this reason, a hidden camera may be installed on a wall, ceiling, object, etc. so that shoplifters cannot easily recognize the surveillance camera.

  Moreover, even if the crime scene can be photographed with the surveillance camera, it is necessary to confirm all the images in order to confirm the images manually, and much labor is required. Therefore, an apparatus capable of automatically detecting a crime scene and recording only the image is desired.

  And if a shoplifter commits a crime, he / she can recognize that the person is a shoplifter and issue a warning notifying that the crime is likely to be done. Therefore, a monitoring system that determines whether a shoplifter is a shoplifter before committing a crime is desired.

  For example, a monitoring system that detects a person's face from video captured by a monitoring camera, counts the number of times the face direction has changed, and if the number of times exceeds a predetermined value, the monitoring system discriminates the person as a shoplifter or the like. Yes (Patent Document 1).

  In addition, there is a monitoring system that detects a person's line-of-sight direction from video captured by a monitoring camera, and discriminates it as a shoplifter or the like based on the time or number of times that the person is looking at a specific direction (Patent Document 2).

  In addition, there is a surveillance camera in which a second camera that captures a blind spot area of the first camera is installed inside the first camera that is visually recognized (Patent Document 3).

JP 2011-4050 A JP 2007-299207 A Japanese Patent No. 5622873

  In Patent Literatures 1 and 2, for example, a person whose line of sight is at a specific position for a long time or the number of times of changing the face direction is increased in order to identify or compare a plurality of products in a department store or the like. Since it is determined as a shoplifter, there is a difficulty in specifying the shoplifter.

  Moreover, in patent document 3, although the crime scene can be image | photographed, if a person does not always confirm a monitor etc. visually, a crime scene cannot be suppressed.

  The present invention has been made in view of the above matters, and the purpose of the present invention is to accurately determine whether the person is a shoplifter before the crime is committed, and to determine the crime scene caused by the shoplifter. The object is to provide a recordable shoplifting crime scene recording device.

Shoplifting crime scene recording device according to the present invention,
A camera that is visible from the outside and can recognize the shooting area or virtual shooting area for shoplifters, and attracts the attention of shoplifters,
A hidden camera for photographing the blind spot area of the alert camera;
Image processing means for processing an image of a person photographed by the hidden camera;
A discriminating means for discriminating whether the photographed person is a shoplifter;
A recording device for recording crime scenes by shoplifters,
The determination means determines the person as a shoplifter when the photographed person performs an action of looking up at the alert camera in the blind spot area of the alert camera,
The recording device records an image of a crime scene photographed by the hidden camera for a predetermined time.
It is characterized by that.

Further, the image processing means extracts a change in the area of the person's face in the captured image,
The discrimination means may discriminate the person as a shoplifter when the area of the face of the person in the photographed image increases.

In addition, the image processing means extracts the line-of-sight direction of the person in the captured image,
The determination means may determine that the person is a shoplifter when the line of sight of the person faces the alert camera.

Further, the image processing means extracts the black eyes of a person in the captured image,
The determination means may determine that the person is a shoplifter when black eyes are extracted.

The image processing means extracts the angle of the person's head in the captured image,
The discriminating means may discriminate the person as a shoplifter when the angle of the head is inclined to the back side.

  The hidden camera may be installed inside the alert camera, and the hidden camera may capture a blind spot area below the alert camera.

  With the shoplifting crime scene recording apparatus according to the present invention, it is possible to accurately determine whether or not the person is a shoplifter before the crime is committed, and to record the crime scene caused by the shoplifting crime.

It is a block diagram which shows schematic structure of a shoplifting crime scene recording device. It is an external appearance perspective view of a warning camera. It is A-A 'sectional drawing of FIG. It is a flowchart which shows operation | movement of a shoplifting crime scene recording device. It is a figure which shows a shoplifter visually recognizing the imaging | photography area | region of an alerting camera and moving. It is a figure which shows a mode that a shoplifter confirms an alerting camera after moving below a alerting camera. It is a graph which shows the example of a time-dependent change of a face area.

  The shoplifting crime scene recording apparatus according to the present embodiment is installed in a store such as a department store, monitors and discriminates shoplifting crimes, and records the crime scene. The shoplifting crime scene recording device is visually recognized from the outside and can recognize the shooting area or the virtual shooting area by the shoplifter, a warning camera that attracts the attention of the shoplifter, and a hidden camera that captures the blind spot area of the warning camera And image processing means for processing an image of a person photographed with a hidden camera, determination means for determining whether or not the photographed person is a shoplifter, and a recording device for recording a crime scene caused by the shoplifter.

  Specifically, as shown in FIG. 1, the shoplifting crime scene recording apparatus 1 includes an alerting camera 10, a hidden camera 20, an image processing device 30, a recording device 40, and a notification device 50.

  As shown in FIGS. 2 and 3, in the present embodiment, the hidden camera 20 is installed inside the alerting camera 10. The alert camera 10 has a dummy lens 12 installed on one surface of a rectangular parallelepiped case 11. The dummy lens 12 is installed so that it can be easily seen from the outside. For this reason, a person who sees the attention-calling camera 10 recognizes that an image is shot through the dummy lens 12. Note that a commercially available dummy surveillance camera may be used as it is as the alert camera 10. For example, a hole 13 of about 1 mm to 5 mm is provided in a part of the side surface of the case 11 of the alert camera 10. The hidden camera 20 takes an image through the hole 13.

  The hidden camera 20 images the blind spot in the virtual shooting area of the alert camera 10. The hidden camera 20 is installed inside the case 11 of the alert camera 10 and is arranged so that it cannot be seen from the appearance of the alert camera 10. The hidden camera 20 is fixed to the case 11 so that the lens of the hidden camera 20 faces the hole 13. The hidden camera 20 is connected to an external control unit 34 by wiring 21. Note that the video may be transmitted by wireless communication without providing the wiring 21.

  Any camera may be used as long as the hidden camera 20 is smaller than the internal volume of the case 11 and can be installed inside the case 11 without being recognized from the appearance of the alert camera 10. For example, a pinhole lens camera can be mentioned. A pinhole lens camera projects an image on an individual imaging element such as a CCD image sensor or a CMOS image sensor disposed behind a pinhole lens through a hole having a small diameter of about 1 to 4 mm, for example. Since photography can be performed through a small-diameter hole, the diameter of the hole 13 provided in the case 11 can be as small as several millimeters. For this reason, the alerting camera 10 is visually recognized in the same manner as a generally used monitoring camera, and therefore it is impossible to recognize from the outside that the hidden camera 20 is installed. Moreover, when the existing screw hole, the ventilation hole, etc. are provided in the case of the commercially available dummy camera, they may be utilized as the hole 13.

  The lens of the hidden camera 20 is installed so as to face a different direction from the dummy lens 12 of the alerting camera 10, for example, as shown in FIG. 3, the optical axis of the dummy lens 12 of the alerting camera 10 and the hidden camera The hidden camera 20 is installed in the case 11 of the alerting camera 10 so that the crossing angle θ with the optical axis of the 20 lenses is 80 to 100 °.

  In the present embodiment, a plurality of alerting cameras 10 and hidden cameras 20 are installed at a plurality of locations in a store such as a department store. And the authentication number is attached | subjected to each set of the alerting camera 10 and the hidden camera 20, and it has come to understand which image was image | photographed with which hidden camera 20. FIG.

  The image processing device 30 includes a memory 31, an image processing unit 32, a determination unit 33, and a control unit 34.

  The memory 31 is a non-volatile memory such as a RAM, and temporarily stores video captured by the hidden camera 20.

  The image processing unit 32 processes an image captured by the hidden camera 20 based on an instruction from the control unit 34.

  Based on an instruction from the control unit 34, the determination unit 33 determines whether a person exists in the blind spot below the alerting camera 10 and whether the person is a shoplifter.

  The control unit 34 controls the entire components of the shoplifting crime scene recording apparatus 1.

  The recording device 40 is a recording device such as a hard disk, and records a photographed image of a crime scene by a shoplifter taken by the hidden camera 20 based on an instruction from the control unit 34. For example, the recording device 40 records a crime image for about one minute after it is determined to be a shoplifter. Also, the authentication number and date / time of the hidden camera 20 that has taken the crime image are registered.

  The notification device 50 is a light emitting device, a speaker, a monitor, or the like, and notifies that a shoplifter who intends to shoplift is below the alerting camera 10 based on an instruction from the control unit 34.

  Next, the operation of the shoplifting crime scene recording apparatus 1 will be described with reference to FIG.

  As shown in FIG. 5, the alerting camera 10 is installed in a place such as a ceiling that is easily visible from a person. For this reason, the virtual imaging area of the dummy lens 12 of the alerting camera 10 is recognized by a person. Then, the hidden camera 20 sequentially captures the blind spot below the alert camera 10. An image taken by the hidden camera 20 is sent to the memory 31.

  The image processing unit 32 continuously processes the images stored in the memory 31. Specifically, whether or not there is a person below the monitoring camera is processed so that the determination unit 33 can detect a moving object. If there is no change in the continuous frames processed by the image processing unit 32 (step S1: NO), the determination unit 33 returns to step S1, assuming that no person exists below the alerting camera 10.

  On the other hand, when there is a change in successive frames, the determination unit 33 determines that a person exists below the alerting camera 10 (step 1: YES).

  The above image processing may be performed by a known method. For example, when an image is binarized to be divided into white and black, and when changing from white to black or from black to white between consecutive frames, a person shall exist below the alerting camera 10. Can do. Also, the image is binarized and divided into white and black, the area of the white or black part is measured, and when the area of the white or black part changes, a person is present below the alerting camera 10. be able to. Further, it may be processed in gray scale or color.

  After determining that a person is present below the alerting camera 10, the image processing unit 32 extracts the face of the person in the image in each frame that is subsequently captured (step S2). Then, the area of the extracted person's face is calculated (step S3). Then, the area change rate of the face is calculated (step S4).

  And the discrimination | determination part 33 discriminate | determines whether the person who is image | photographed looked up at the alerting camera 10, ie, it is a shoplifter who is going to perform shoplifting from now on (step S5). When the person being photographed performs an operation of looking up at the alerting camera 10 (step S5: YES), it is determined that the person is a shoplifter.

  Usually, the shoplifter S picks up a product, recognizes the alert camera 10 and its virtual shooting area, and heads to the blind spot of the alert camera 10. As shown in FIG. 5, the normal shoplifter S recognizes that if it is a blind spot below the alert camera 10, it is a place where the crime can be safely committed without being photographed. Then, the shoplifter S who has moved below the alert camera 10 performs an operation of looking up at the alert camera 10 as shown in FIG. 6 in order to confirm that it is a blind spot in the virtual shooting area of the alert camera 10. . At this time, the face of the shoplifter S faces the alert camera 10. When the shoplifter S performs an action of looking up at the alert camera 10 (step S5: YES), the determination unit 33 determines that the person is a shoplifter S who is about to commit a crime.

  The calculation of the face area is, for example, a known image processing method such as a method of calculating an area (number of dots) converted to white by the above-described binarization processing of an image obtained by extracting a human face. The face area of a person in the image can be calculated.

  FIG. 7 shows an example of the change over time of the face area in the image taken by the hidden camera 20. In the case of a general customer, even if the user is below the alerting camera 10, the user is not interested in the alerting camera 10. For this reason, in the case of a general customer, the change in the area of the face is small even if time passes.

  On the other hand, in the case of a shoplifter who intends to shoplift, after moving to the lower side of the alert camera 10, an operation of looking up at the alert camera 10 is performed to confirm that the camera is not photographed. For this reason, in the case of a shoplifter, a part where the area of the face in the photographed image increases suddenly occurs. When the area change rate of the face increases, the determination unit 33 determines the person as a shoplifter.

  For example, when confirming that a shoplifter is a blind spot area of the alert camera 10, an operation of looking up for a moment is often performed. For this reason, a face area change rate threshold value (for example, a value of a general customer face area change rate) per unit time (for example, per 0.2 seconds) is set in advance, and the face area change exceeding the set threshold value is set. If the rate is detected, the person may be identified as a shoplifter.

  When the determination unit 33 determines that the person is a shoplifter, the recording device 40 records a captured image of the hidden camera 20 for a predetermined time (for example, one minute) based on an instruction from the control unit 34 (step S6). Recording may be continued until there is no change in consecutive frames of the captured image, that is, until shoplifter S no longer exists from below the alerting camera 10. As a result, when the shoplifter S commits a crime, the entire crime scene is recorded in the recording device 40 as an evidence image.

  Further, when the determination unit 33 determines that the shoplifter S is a shoplifter S, the notification device 50 has a shoplifter S below the alerting camera 10 based on an instruction from the control unit 34, and it is highly probable that a crime will be performed in the future. This is notified to the surroundings (step S7). The notification is performed by, for example, sound or light, display on a display, or the like.

  When it is determined that the person has not looked up at the alerting camera 10 (step S5: NO), the process returns to step S1.

  In addition, although the example which discriminate | determines whether the operation | movement which looks up at the alerting camera 10 was performed based on the change of the area of a face was demonstrated above, you may comprise so that it may discriminate | determine by another method.

  For example, when two black eyes of a person photographed by the hidden camera 20 are reflected in the image processing, the person may be identified as the shoplifter S. In the case of a shoplifter who intends to shoplift, in order to confirm that it is a blind spot area of the alert camera 10, the image taken by the hidden camera 20 has 2 black eyes of the shoplifter in order to look up at the alert camera 10. Will be projected.

  Alternatively, in the image processing, information on the gaze direction of the person photographed by the hidden camera 20 is acquired, and when the gaze direction of the person faces the direction of the alerting camera 10, the person is identified as the shoplifter S. Form may be sufficient. These can be performed by a known gaze direction measurement method or the like.

  In addition, the contour around the person's face is extracted, and whether or not the person looks up at the alert camera from the change in contour movement (for example, whether the head angle is tilted toward the back side by extracting the change in the head angle) For example, various known operation determination methods using processing of an image taken by the hidden camera 20 may be used.

  In the above description, the example in which the hidden camera 20 is built in the alert camera 10 has been described. However, it is only necessary that the shoplifter can determine the action of looking up at the alert camera 10 from the captured image. It may be installed on a ceiling, a product shelf, a wall, etc. around the alerting camera 10 and may take an image of a blind spot area below the alerting camera 10.

  Further, the example in which the alerting camera 10 is a dummy camera has been described, but a normal camera capable of photographing may be used.

  As described above, the shoplifting crime scene recording device 1 of the present application skillfully uses the psychology of the shoplifter S, hides the shoplifter S and squeezes it into the shooting area of the camera 20, and the behavior that shoplifters perform before the crime. By using this, it is possible to discriminate in advance the shoplifter who has a high probability of committing a crime. Therefore, it is possible to accurately monitor whether or not the person is a shoplifter before committing a crime, and the effectiveness is high. And since the recording device 40 has recorded the crime scene picked up from the image | video image | photographed with the hidden camera 20 for the predetermined time, even if it does not see all the images | videos imaged with the hidden camera 20, it is a shoplifting crime scene. This is a highly effective device with a wide range of methods of use, such as submission of crime scene images to the police and use of warning data in the store.

1 Shoplifting Crime Scene Recording Device 10 Warning Camera 11 Case 12 Dummy Lens 13 Hole 20 Hidden Camera 21 Wiring 30 Image Processing Device 31 Memory 32 Image Processing Unit (Image Processing Means)
33 Discrimination part (discrimination means)
34 Control unit 40 Recording device 50 Notification device S Shoplifter

Claims (6)

A camera that is visible from the outside and can recognize the shooting area or virtual shooting area for shoplifters, and attracts the attention of shoplifters,
A hidden camera for photographing the blind spot area of the alert camera;
Image processing means for processing an image of a person photographed by the hidden camera;
A discriminating means for discriminating whether the photographed person is a shoplifter;
A recording device for recording crime scenes by shoplifters,
The determination means determines the person as a shoplifter when the photographed person performs an action of looking up at the alert camera in the blind spot area of the alert camera,
The recording device records an image of a crime scene photographed by the hidden camera for a predetermined time.
A shoplifting crime scene recording device characterized by that. The image processing means extracts a change in the area of the person's face in the captured image,
The determining means determines the person as a shoplifter when the area of the face of the person in the captured image increases.
The shoplifting crime scene recording apparatus according to claim 1. The image processing means extracts the line-of-sight direction of the person in the captured image,
The determination means determines the person as a shoplifter when the line of sight of the person faces the alert camera.
The shoplifting crime scene recording apparatus according to claim 1. The image processing means extracts the black eyes of a person in the captured image,
The determining means determines that the person is a shoplifter when black eyes are extracted;
The shoplifting crime scene recording apparatus according to claim 1. The image processing means extracts the angle of the person's head in the captured image,
The discrimination means discriminates the person as a shoplifter when the angle of the head is inclined to the back side,
The shoplifting crime scene recording apparatus according to claim 1. The hidden camera is installed inside the alerting camera, and the hidden camera captures a blind spot area below the alerting camera;
The shoplifting crime scene recording apparatus according to any one of claims 1 to 5, wherein the shoplifting crime scene recording apparatus is provided.

Images