JP2012103901A - Intrusion object detection device - Google Patents

Abstract

Provided is an intruding object detection device capable of specifying a detection area for detecting an intruding object based on a thermal change on a monitoring image with high accuracy and improving the detection accuracy of the intruding object.
An intruding object detection apparatus includes a sensor unit that outputs a detection level corresponding to heat rays from a detection area, an imaging unit that acquires a monitoring image, and a change region extraction unit that extracts a change region. The human image position extracting means 73 extracts the position of the change area extracted from the monitoring image when the detection level reaches a predetermined value as the human image position, and based on the plurality of human image positions extracted in the predetermined period, An area specifying unit 74 that detects an end of the moving range on the monitoring image and specifies a sensor area based on the end, and an abnormality that differs depending on whether the change area exists outside the sensor area or the sensor area It has an abnormality monitoring means 75 for executing the determination process.
[Selection] Figure 1

Description

  The present invention relates to an intruding object detection device that detects an intruding object based on an image obtained by capturing a thermal change in a detection area and a monitoring area including the detection area.

  In recent years, a technique for detecting an intruding object using different types of sensors has been developed in order to detect an intruding object entering the monitoring area with high accuracy. For example, Patent Document 1 discloses a composite monitoring apparatus that uses a pyroelectric infrared sensor (PIR sensor; Passive InfraRed Sensor) and an image sensor as sensors for monitoring an intruding object. The composite monitoring apparatus disclosed in Patent Document 1 determines the presence or absence of an intruding object by comprehensively determining the detection results of two types of sensors.

The composite monitoring apparatus disclosed in Patent Document 1 is based on the premise that the detection area where the PIR sensor detects an intruding object and the imaging range captured by the image sensor are equal. However, in such a composite type monitoring apparatus, it may be preferable to change the detection area of the PIR sensor and the imaging range of the image sensor according to the monitoring purpose, installation environment, and the like.
For example, by setting the imaging range of the image sensor wider than the detection area of the PIR sensor, when an intruding object is detected, the supervisor can browse the monitoring image captured by the image sensor and report the cause of the alarm and the surrounding situation. Can be easily grasped. Further, in the composite type monitoring device, when a factor that causes the PIR sensor to erroneously detect an intruding object (such as an object whose temperature changes frequently) is included in the monitoring area, the factor is included in the detection area of the PIR sensor. It is necessary to narrow the detection area of the PIR sensor with respect to the imaging range of the image sensor. In some cases, it is preferable that only the portion of the monitoring area to be monitored with priority is set as the detection area of the PIR sensor.

  In this way, when the imaging range of the image sensor and the detection area of the PIR sensor are different, the composite monitoring apparatus detects the imaging range of the image sensor and the detection of the PIR sensor in order to associate the detection result by the image sensor with the detection result by the PIR sensor. It is necessary to recognize the relationship between areas. Therefore, for example, when the detection area is part of the shooting range, the composite monitoring apparatus needs to specify the position (range) of the detection area on the monitoring image. However, particularly when the orientation of the camera of the image sensor and the orientation of the PIR sensor can be arbitrarily changed, the range of the detection area on the monitoring image changes depending on the relative relationship between the orientation of the camera and the orientation of the PIR sensor. It is difficult to specify the range of the detection area on the monitoring image.

  On the other hand, Patent Literature 2 discloses a technique for specifying the range of the detection area of the PIR sensor on the monitoring image. The monitoring image processing device disclosed in Patent Document 2 extracts the position of a moving object (operator) on a monitoring image taken when a walk test is performed by an operator by image processing of motion detection, A set of positions of moving objects on the monitoring image when the PIR sensor detects an object is set as a detection area.

JP 2000-33252 A JP 2006-178515 A

In general, when using a PIR sensor, a detection area composed of one or a plurality of detection zones is set, and the PIR sensor detects a change in thermal energy in each detection zone using a pyroelectric element. Due to such characteristics, the PIR sensor requires a certain amount of time accompanying the movement of the human body in order to detect a change in thermal energy, and therefore the position of the human body when the PIR sensor detects a change in thermal energy, There may be a difference between the detection area (detection zone) where the human body actually exists.
Therefore, in the technique described in Patent Document 2, the range of the detection area of the PIR sensor on the monitoring image may not be correctly specified. Therefore, even if the technique described in Patent Document 2 is applied to a composite monitoring apparatus that uses both an image sensor and a PIR sensor, the relationship between the imaging range of the image sensor and the detection area of the PIR sensor may not be recognized correctly. is there. In this case, the composite monitoring apparatus cannot associate the detection result by the image sensor with the detection result by the PIR sensor, and cannot sufficiently improve the detection accuracy of the intruding object.

  Accordingly, an object of the present invention is to provide an area representing a detection area on a monitoring image in an intruding object detection device that detects an intruding object based on a monitoring image obtained by capturing a thermal change in the detection area and a monitoring area including the detection area. It is an object of the present invention to provide an intruding object detection apparatus that can identify the intruding object with high accuracy and improve the intrusion object detection accuracy.

  The present invention for solving such a problem includes a sensor unit that outputs a detection level corresponding to a heat ray radiated from a detection area, a limit object and a detection area, in the vicinity of a restricted object that restricts movement of a person. An intruding object detection apparatus is provided that includes an imaging unit that acquires a monitoring image obtained by imaging a monitoring area including a predetermined period. Such an intruding object detection device includes a change area extraction unit that extracts a change area in a monitoring image, an area setting unit that executes a registration process for specifying a sensor area indicating a detection area on the monitoring image, a detection level, and a monitoring image. An abnormality monitoring means for executing an abnormality determination process for detecting an intruding object in the monitoring area based on the detection level output by the sensor unit to a value indicating that a person is present in the detection area. A human image position extracting unit that extracts the position of the change area extracted from the monitoring image acquired by the imaging unit as a human image position, and a plurality of human image positions extracted by the human image position extracting unit during a predetermined period. Based on the end of the movement range of the person limited by the restriction object is detected on the monitoring image, and the restriction of the sensor area based on the end of the movement range And an area specifying means for specifying a sensor area by defining a first end that is an end on the side, and the abnormality monitoring means depends on whether the change area exists outside the sensor area or outside the sensor area Different abnormality determination processing is executed.

  Further, in the intruding object detection device according to the present invention, the appearance frequency of the human image position extracted during a predetermined period is obtained for each position on the monitoring image, and the position where the appearance frequency rapidly decreases with respect to the adjacent position is moved. The end of the range is preferred.

  Alternatively, in the intruding object detection device according to the present invention, the human image position extraction means stores the human image position and the detection level in association with each other when the detection level becomes a value indicating that a person is present in the detection area, It is preferable that the specifying unit sets the position where the detection level stored in association with the human image position on the monitoring image rapidly decreases with respect to the adjacent position as the end of the movement range.

  Furthermore, in the intruding object detection device according to the present invention, the abnormality monitoring means may be configured such that when the newly extracted change area exists in the vicinity of the first end of the sensor area, the change area is not caused by the intruding object. It is preferable to determine that there is no.

  An intruding object detection device that detects an intruding object based on a thermal change in a detection area according to the present invention and a monitoring image obtained by imaging a monitoring area including the detection area, the area representing the detection area on the monitoring image with high accuracy. In particular, it is possible to improve the detection accuracy of the intruding object.

It is a schematic block diagram which shows the whole system structure of the security system to which this invention is applied. (A) is the figure which looked at the imaging range of the imaging part, and the detection area of the sensor part from the diagonal direction, (b) is the figure which looked at the imaging range of the imaging part, and the detection area of the sensor part from the perpendicular direction. (C) is a figure showing the monitoring image of an imaging part, and the detection area of a sensor part. It is an example of appearance frequency distribution of a human image position. It is a flowchart which shows the process at the time of the registration mode by an area setting means. It is a flowchart which shows the process at the time of warning mode by a change area extraction means and an abnormality monitoring means. It is an example of peak distribution of a detection level. It is a flowchart which shows the process at the time of the registration mode by an area setting means.

Hereinafter, an intruding object detection apparatus according to an embodiment of the present invention will be described with reference to the drawings.
The intruding object detection device detects an intruding object based on a thermal change in the detection area, and acquires a monitoring image obtained by photographing a monitoring area including the detection area. The detection area is an area adjacent to an object (such as a structure such as a building wall) that restricts the movement of a person, and the monitoring image includes a part or all of the restricted object. In particular, in this intruding object detection device, a walk test (test walking) is performed in which the worker walks in the monitoring area to confirm the position of the detection area on the monitoring image. At this time, the intruding object detection device extracts the position of the human body from the monitoring image acquired when it is determined that the human body exists due to the thermal change in the detection area. Then, the intruding object detection device specifies the boundary of the detection area on the monitoring image based on the distribution of the appearance frequency of the position of the human body. Furthermore, the intruding object detection device determines the presence or absence of an intruding object by a different determination process between the area representing the detection area on the monitoring image and the other area. As a result, the intruding object detection device can accurately distinguish between the area representing the detection area on the monitoring image and the other area, and can determine the presence or absence of the intruding object in each area by appropriate determination processing. The detection accuracy of the intruding object can be improved.

  FIG. 1 is a diagram showing an overall system configuration of a security system to which the present invention is applied. As shown in FIG. 1, the security system 100 includes intruding object detection devices 1 and 2, a security device 3, an operation indicator 4, and a monitoring center device 5. The intruding object detection devices 1 and 2, the security device 3, and the operation indicator 4 are installed in the monitoring area or in the vicinity thereof, and are connected by a wireless or wired local area network 6. On the other hand, the monitoring center device 5 is installed in a remote security center and is connected to the security device 3 via a wide area communication network 7 such as a public telephone line.

  When the intruding object detection devices 1 and 2 detect an intruding object that has entered the monitoring area, the intruding object detection signal indicating that the intruding object has been detected and a monitoring image obtained by capturing the monitoring area at that time are displayed on the local area network 6. To the security device 3. When the security device 3 receives the intruding object detection signal and the monitoring image from the intruding object detection device 1 or 2, the security device 3 monitors information such as the intruding object detection signal, the monitoring image, and an identification code that identifies the monitoring area where the intruding object is detected. Transmit to the center device 5. Note that the identification code for specifying the monitoring area where the intruding object is detected can be, for example, the identification code of the intruding object detecting devices 1 and 2, the security device 3, or the corresponding monitoring area.

  Further, the intruding object detection devices 1 and 2 may periodically transmit the monitoring image to the operation display 4. The operation indicator 4 has, for example, a touch panel display, and displays operation buttons for the user to input operation commands. When the user presses one of the displayed operation buttons, the operation display 4 displays a monitoring image received from the intruding object detection device 1 or 2 according to an operation command corresponding to the operation button. Note that the operation indicator 4 is not a touch panel display, and may have a display device for displaying information and an input device for a user to perform an input operation such as a keypad or a trackball. Good.

The monitoring area may be outdoors or indoors, but in the present embodiment, a case where a garden of a house is monitored as the monitoring area will be described as an example.
Further, the intruding object detection devices 1 and 2 and the operation indicator 4 may be configured to be connected to the security device 3 through signal lines and to perform communication between the devices via the security device 3.
Further, when the security device 3 receives the intruding object detection signal from the intruding object detection device 1 or 2, the security device 3 determines whether or not there is an intrusion abnormality and then transmits information such as the intruding object detection signal, the monitoring image, and the identification code to the monitoring center device. You may make it transmit to 5. In this case, for example, a reader for detecting an IC tag carried by a user of the security system is provided in the security device 3 separately. The security device 3 determines that there is an intrusion abnormality only when the intruding object detection signal and the monitoring image are received from the intruding object detection device 1 or 2 when the reader does not detect the IC tag in the monitoring area.

  Hereinafter, the configuration and operation of the intruding object detection devices 1 and 2 will be described. Since the configuration and operation of the intruding object detection devices 1 and 2 are the same, the intruding object detection device 1 will be described below as a representative.

  The intruding object detection device 1 is installed on, for example, a wall surface on the outdoor side of a house, and monitors a wall side and a surrounding space including a window or a door with emitted heat rays and captured images. That is, the intruding object detection device 1 detects an invasion action into the indoor portion of the house to be protected in advance on the outdoor side, and accordingly, along the wall (restricted object) that forms a boundary with the indoor portion. Monitor wall space. The intruding object detection device 1 operates in one of two operation modes: a registration mode for setting a monitoring area and the like, and a warning mode for monitoring an intruder who attempts to enter through a window or door. For this purpose, the intruding object detection apparatus 1 includes an imaging unit 10, a sensor unit 20, a storage unit 30, an operation unit 40, a display unit 50, a communication unit 60, and a monitoring processing unit 70.

The imaging unit 10 acquires a monitoring image obtained by capturing a monitoring area including a detection area where the sensor unit 20 detects an intruding object. For this purpose, the imaging unit 10 includes a two-dimensional detector composed of a photoelectric converter such as a CCD or C-MOS sensor, an imaging optical system that forms an image of a monitoring region on the two-dimensional detector, The electric signal output from the two-dimensional detector is amplified and configured by an electric circuit for analog / digital (A / D) conversion. Further, the imaging unit 10 may be an NTSC type camera or a high resolution type digital camera for high definition. The angle of view of the lens of the imaging unit 10 of the present embodiment is 90 ° to 180 °, and the imaging unit 10 is installed so as to be able to photograph a wall surface provided with a window or a door and a space near a wall that is a detection area. The imaging unit 10 captures images at regular time intervals (for example, 1/5 second) to acquire a monitoring image, and transmits the monitoring image to the monitoring processing unit 70.
The monitoring image has, for example, horizontal 340 pixels × vertical 240 pixels, and each pixel is expressed as digital image data having pixel values of 0 to 255. Note that the imaging unit 10 may create the monitoring image as a color image.

The sensor unit 20 includes a pyroelectric infrared sensor (PIR sensor; Passive InfraRed Sensor) using a pyroelectric element and an optical system. A detection area where the PIR sensor detects an intruding object is configured by one or a plurality of detection zones. Each detection zone is composed of a pair of zones, and each detection zone is defined in a region where an intruding object is to be detected.
The PIR sensor includes, for example, a pair of pyroelectric elements connected with opposite polarities. And a PIR sensor condenses the infrared rays (heat ray) from one zone among a pair of zones of each detection zone, receives it with one pyroelectric element, and converts it into a positive voltage. In addition, the PIR sensor collects infrared rays (heat rays) from the other zone of each detection zone by an optical system, receives it by the other pyroelectric element, and converts it into a negative voltage. Therefore, when an intruding object passes in a direction across a pair of zones, a voltage change from a positive voltage to a negative voltage (or from a negative voltage to a positive voltage) occurs in the PIR sensor.
Therefore, the sensor unit 20 obtains a detection level P representing the degree of certainty that the intruding object exists in the detection area based on the magnitude of the voltage change. The detection level P increases as the voltage change increases. For example, the magnitude of the voltage change itself may be used as the detection level P. Then, the sensor unit 20 transmits detection level information indicating the detection level P to the monitoring processing unit 70.

When the intruding object detection device 1 monitors the garden (outdoor) of a house by an image, there is a possibility that a flying bag such as a garbage bag or planting may be erroneously detected as an intruding object. Need to be strict. On the other hand, the intruding object detection device 1 needs to reliably detect an intruder approaching the wall of the house. Therefore, the intruding object detection device 1 according to the present embodiment monitors an intruding object over a wide outdoor area based on the monitoring image, and uses a PIR sensor in combination with a window or door area with a high risk of intrusion. And focus on monitoring intruding objects.
That is, in this embodiment, the space near the wall of the house is set as a detection area of the sensor unit 20, and an area including the detection area is set as a monitoring area captured by the imaging unit 10. Therefore, the viewing angle of the sensor unit 20 may be narrower than the angle of view of the imaging unit 10, for example, about 20 °.

  FIG. 2A is a diagram in which the imaging range of the imaging unit 10 and the detection area of the sensor unit 20 are viewed from an oblique direction. In FIG. 2A, an angle of view 201 is an angle of view of the imaging unit 10 and is set so that the imaging unit 10 captures the imaging range 200. The viewing angle 211 is a viewing angle of the sensor unit 20 and is set so that the sensor unit 20 detects an intruding object in the area of the detection area 210. As shown in FIG. 2A, in the present embodiment, the detection area 210 of the sensor unit 20 is so monitored as to be monitored by the PIR sensor among the areas to monitor the intruding object based on the monitoring image. Is set to be part of the shooting range 200 of the imaging unit 10.

  FIG. 2B is a diagram of the imaging range of the imaging unit 10 and the detection area of the sensor unit 20 as seen from the vertical direction. In FIG. 2B, an angle of view 212 is an angle of view of the imaging unit 10 and is set so that the imaging unit 10 captures the imaging range 202. The viewing angle 213 is the viewing angle of the sensor unit 20 and is set so that the sensor unit 20 detects an intruding object in the area of the detection area 203. As shown in FIG. 2 (b), the intruding object detection device 1 is installed on the wall 220 (vertical line portion) of the house so that an intruding object entering the house can be detected and photographed. The shooting range 202 of the imaging unit 10 and the detection area 203 of the sensor unit 20 are set so as to include the wall, respectively, and the shooting range 202 of the shooting unit 10 is set to include the wall surface of the wall 220.

  FIG. 2C is a diagram illustrating a monitoring image of the imaging unit 10 and a detection area of the sensor unit 20. In FIG. 2C, an image 204 is a monitoring image captured by the imaging unit 10, and a region 214 (hatched portion) in the monitoring image 204 represents a detection area of the sensor unit 20, and a region 221 (vertical line portion). ) Represents an area where the wall 220 of the house was imaged. As shown in FIG. 2C, the imaging unit 10 images a space along the house wall 220, and the sensor unit 20 uses a region 214 along the house wall 220 in the monitoring image 204 as a detection area. The monitoring image 204 is set so as to include the detection area 214.

  As described above, in the intruding object detection device 1 according to the present embodiment, since the imaging range of the imaging unit 10 is set so as to include the detection area of the sensor unit 20, when the detection area is set near the wall of a house, monitoring is performed. The wall surface of the image is captured on either the left or right end side. A region having a predetermined size determined from the wall surface by the viewing angle of the sensor unit 20 and the angle of view of the imaging unit 10 represents a detection area of the PIR sensor on the monitoring image (hereinafter referred to as a sensor region). It becomes.

  In the present embodiment, the sensor unit 20 has a pan function and can rotate in the horizontal (left and right) direction with respect to the visual axis of the imaging unit 10. That is, the position of the sensor area is fixed in the vertical (vertical) direction and varies in the horizontal (horizontal) direction. The horizontal orientations of the sensor unit 20 and the imaging unit 10 are adjusted manually by the installer according to the installation environment of the intruding object detection device 1, but at least the imaging range of the imaging unit 10 is within the sensor unit 20. It is set to include the detection area. The imaging unit 10 may also be provided with a pan function so that the visual axis can be rotated in the horizontal direction with respect to the intruding object detection device 1. In addition, the sensor unit 20 and the imaging unit 10 may further include a tilt function, and the respective visual axes may be rotated in the vertical direction with respect to the intruding object detection device 1.

  The storage unit 30 includes a nonvolatile semiconductor memory such as a flash memory (registered trademark), a volatile semiconductor memory, a magnetic disk (HDD), an optical disk drive such as a CD-ROM and a DVD-RAM, and a recording medium thereof. The storage unit 30 stores a program executed on the monitoring processing unit 70, various setting parameters used by the program, a calculation value or an image obtained as a result of various processes or an intermediate result. Then, the storage unit 30 stores the calculated value or the image according to the control signal from the monitoring processing unit 70 or outputs various stored information to the monitoring processing unit 70.

  The operation unit 40 includes input devices such as a keyboard and a mouse, receives an operation for switching between a registration mode and a warning mode or various operations in the registration mode from an operator, and sends signals corresponding to those operations to the monitoring processing unit 70. Output.

  The display unit 50 is configured by a display device such as a liquid crystal display, displays various information received from the monitoring processing unit 70, and notifies the operator.

  The input of various operations and the display of various information may be realized by a terminal different from the intruding object detection device 1 such as a PDA (Personal Digital Assistants). In that case, the operation unit 40 and the display unit 50 communicate with a terminal such as a PDA via the communication unit 60 to input various operations and display various information.

  The communication unit 60 is an input / output interface that transmits and receives various setting signals and control signals between the intruding object detection device 1 and the security device 3 via a communication network such as the local area network 6. ) And the like, and driver software for driving them.

  The monitoring processing unit 70 has an embedded microprocessor unit and its peripheral circuits, and controls the entire intruding object detection apparatus 1. In the registration mode, the monitoring processing unit 70 identifies the sensor area on the monitoring image received from the imaging unit 10, and in the monitoring mode, the monitoring image received from the imaging unit 10 and the detection level information received from the sensor unit 20 are used. Based on this, an intruding object in the monitoring area is detected and an alarm is issued. For this purpose, the monitoring processing unit 70 includes a change area extraction unit 71, an area setting unit 72, and an abnormality monitoring unit 76. Each of these units included in the monitoring processing unit 70 is implemented as, for example, a functional module of a program that operates on the microprocessor unit.

  The change area extraction unit 71 extracts an area where a moving object is captured from the monitoring image every time the imaging unit 10 acquires the monitoring image. For example, the change area extraction unit 71 extracts a change area having a temporal change in luminance value from a plurality of monitoring images continuously acquired by the imaging unit 10 using inter-frame difference processing or background difference processing. To do. Then, the change area extraction unit 71 stores the extracted change area in the storage unit 30. The change area is an area where pixels having a change in luminance of a predetermined level or more are gathered to some extent, and a change in noise level or a change in only a minimum area is removed.

  The area setting unit 72 specifies a sensor region based on the detection level information acquired from the sensor unit 20 and the monitoring image acquired from the imaging unit 10 when the worker is performing the walk test in the registration mode. For this purpose, the area setting unit 72 includes a human image position extracting unit 73, an area specifying unit 74, and an area registration unit 75.

  The area setting unit 72 determines whether or not a walk test is being performed by receiving a signal indicating that a walk test start operation or end operation has been performed from the operation unit 40. Alternatively, the walk test may be automatically ended after a predetermined period of time has elapsed after receiving a signal indicating the start operation of the walk test.

The human image position extraction means 73 receives detection level information from the sensor unit 20 during the walk test, that is, when the worker is walking in the monitoring area, and the detection level P is equal to or higher than a predetermined reference value Z1. It is determined whether or not. Then, when the detection level P is equal to or higher than the reference value Z1, the human image position extracting unit 73 extracts the center of gravity position of the changed area extracted by the changed area extracting unit 71 at that time as the human image position and stores it in the storage unit 30.
In addition, the reference value Z1 is set to a value indicating that the detection level P has a person in the detection area. Also, in the walk test, it is preferable to make it easier to detect people than in the alert mode so that many samples can be acquired in a short time, and therefore the reference value Z1 is a threshold for determining an intruding object in the alert mode. The following values are preferable.

  Further, in the intruding object apparatus 1, since the sensor unit 20 and the imaging unit 10 operate at independent timings, the timing at which the sensor unit 20 outputs detection level information to the monitoring processing unit 70 and the imaging unit 10 acquire a monitoring image. Timing does not necessarily match. Therefore, the timing at which the human image position extraction unit 73 determines the detection level does not necessarily match the timing at which the change region extraction unit 71 acquires the image for extracting the change region. Therefore, the human image position extracting unit 73 extracts the human image position from the change area of the monitoring image immediately before it is determined that the detection level P is equal to or higher than the reference value Z1. Alternatively, the human image position extracting unit 73 extracts the human image position from the change area of the monitoring image immediately after determining that the detection level P is equal to or higher than the reference value Z1.

After the end of the walk test, the area specifying unit 74 analyzes the position of the human image stored in the storage unit 30 during the walk test, and determines the end of the movement range of the person restricted by the restriction that restricts the movement of the person. Detect on the monitoring image. The area specifying unit 74 specifies the position of the restriction object on the monitoring image from the end of the detected movement range of the person. In this embodiment, the wall of a house is assumed as a restriction | limiting object, and the area specification means 74 specifies the wall surface position which is a position where the wall surface was copied as a position of a restriction | limiting object. As shown in FIGS. 2B and 2C, in this embodiment, the imaging unit 10 is installed so as to capture the wall 220 of the house from the horizontal direction and the monitoring area from the horizontal direction. Therefore, the horizontal position at which the wall surface on the monitoring image is imaged does not differ greatly between the vicinity of the imaging unit 10 (lower side on the monitoring image) and the far side (upper side on the monitoring image). Therefore, the area specifying unit 74 specifies the wall surface position based on the horizontal position on the monitoring image.
As shown in FIGS. 2 (b) and 2 (c), when monitoring a wall, a person's movement range is limited by a wall surface at one end and not at the other end. Therefore, when a test walk is performed within the shooting range, the human image position when the person is detected by the sensor unit 20 is not likely to deviate from the actual human position on the wall side, and the movement is not hindered on the opposite side. Is likely to occur. Therefore, the area specifying unit 74 detects which end in the horizontal direction is on the wall side with respect to the movement range of the person during the walk test.

  In the present embodiment, the area specifying unit 74 first creates an appearance frequency distribution at each horizontal position on the monitoring image for the human image position stored in the storage unit 30. At this time, in order to reduce the influence of the measurement error, it is preferable that the area specifying unit 74 obtains the number of appearances of the human image position for each certain range. For example, the area specifying unit 74 divides the monitoring image into a plurality of areas having a predetermined width in the horizontal direction, and obtains the number of times the human image position is recorded for each divided area. Alternatively, the area specifying unit 74 may obtain the number of times the human image position is recorded in an area having a predetermined width centered on the horizontal position for each horizontal position on the monitoring image, and the number of appearances corresponding to each horizontal position may be used. .

  FIG. 3 shows an example of the appearance frequency distribution of human image positions. In FIG. 3, a graph 300 represents the appearance frequency distribution of the human image position on the monitoring image, the horizontal axis represents the horizontal position on the monitoring image, and the vertical axis represents the number of appearances of the human image position at each position. In the graph 300, the number of appearances of the human image position has a certain value or more in a certain range on the left side of the coordinate X1, but rapidly decreases at the coordinate X1, and becomes zero in the range on the right side of the coordinate X1. That is, in this case, it can be estimated that the worker has not moved to a location corresponding to the position on the right side of the coordinate X1, and the wall surface position X2 exists in the vicinity thereof.

Therefore, the area specifying unit 74 obtains a position (hereinafter referred to as an edge position) where the number of appearances rapidly decreases in the appearance frequency distribution of the created human image position. For example, the area specifying unit 74 sets a position where the number of appearances of the human image position is equal to or less than a predetermined number and decreases by a predetermined threshold or more with respect to the number of appearances of the human image position in the adjacent position as the edge position. Furthermore, it is preferable that the number of appearances at subsequent horizontal positions be 0 or a predetermined number or less. Alternatively, the area specifying unit 74 generates an approximate curve of the number of appearances of the human image position using a least square method or the like, and the value of the approximate curve at each horizontal position is equal to or less than a predetermined value, and the approximate curve at the adjacent position is used. A position that decreases by a predetermined threshold or more with respect to the value may be set as the edge position. Further, it may be added to the condition for determining the edge position that the appearance frequency continues at 0 or below a predetermined horizontal position after the position where the appearance frequency has rapidly decreased. It should be noted that these predetermined threshold values relating to the reduction range can be determined appropriately by, for example, performing a walk test by experiment and measuring the number of appearances of the human image position. Further, in this example, not only the reduction width but also the absolute number (value) at the horizontal position is not more than a predetermined value as a condition for the edge position. This is because the appearance frequency is 0 or a number (value) very close to 0 on the side where movement is restricted by the wall.
Note that, as shown in FIGS. 2B and 2C, in general, the imaging unit 10 covers a wider area on the side opposite to the wall 220 of the house so that the monitor can grasp the surrounding situation in more detail by the image. It is installed so that it may image. That is, when a wall surface is imaged in the right direction, the wall surface is imaged on the right side from the center position on the monitoring image, and when a wall surface is imaged on the left side, the wall surface is imaged on the left side from the center position on the monitoring image. The Therefore, the area specifying unit 74 extracts an edge position based on the amount of decrease in the number of appearances of the human image position relative to the position adjacent to the left side in the region on the right side of the central position on the monitoring image. In the left area, the edge position may be extracted based on the amount of decrease in the number of appearances of the human image position relative to the position adjacent to the right side.

  Then, the area specifying unit 74 estimates that the obtained edge position is a wall surface position. In consideration of the fact that the worker does not move to a position in contact with the wall surface, the area specifying means 74 is the side where the number of appearances of the human image position is smaller than the obtained edge position, that is, the opposite side of the movement range of the person (see FIG. In the example shown in FIG. 3, the position shifted by a predetermined amount on the right side) may be estimated as the wall surface position. Note that this predetermined amount can also be determined as an appropriate value by, for example, performing a walk test by experiment and measuring the difference between the edge position and the wall surface position.

  Then, the area specifying means 74 uses the estimated wall surface position as the end of the sensor region on the restricted object side, and from the wall surface position, the side where the number of appearances of the human image position is large, that is, the human movement range side ( On the left side, an area having a size determined based on the viewing angle of the sensor unit 20, the angle of view of the imaging unit 10, and the like is specified as the sensor area. In the example shown in FIG. 3, the shaded portion 301 is specified as the sensor region.

  When the edge position cannot be detected after the walk test is performed, the detection area of the sensor unit 20 is highly likely not to include an area along the wall surface. Therefore, if the edge position cannot be detected, the area specifying unit 74 displays on the display unit 50 that the detection area of the sensor unit 20 has not been set correctly and notifies the operator. Thereby, the intruding object detection device 1 can prevent an artificial setting error regarding the direction of the PIR sensor, and can improve the reliability of intrusion monitoring.

  The area registration means 75 generates sensor area information indicating the position of the sensor area specified by the area specifying means 74 and information outside the sensor area indicating the position of an area other than the sensor area, and registers them in the storage unit 30.

  Hereinafter, the processing in the registration mode controlled by the area setting means 72 will be described with reference to the flowchart shown in FIG. In the processing described below, the operator uses the operation unit 40 to perform the switching operation from the warning mode to the registration mode and the start operation of the walk test, and the area setting means 72 is a signal corresponding to these operations. Is executed at the timing of receiving from the operation unit 40.

  The worker moves across the entire monitoring area after performing the start operation of the walk test using the operation unit 40. At this time, the human image position extraction unit 73 receives the detection level information from the sensor unit 20, and determines whether or not the detection level P is equal to or higher than the reference value Z1 (step S1). When the detection level P is less than the reference value Z1, the human image position extraction unit 73 returns the control to step S1 and waits until new detection level information is received from the sensor unit 20.

  On the other hand, when the detection level P becomes equal to or higher than the reference value Z1, the human image position extraction unit 73 extracts the center of gravity position of the change area extracted by the change area extraction unit 71 at that time as the human image position and stores it in the storage unit 30 ( Step S2). Thereafter, the human image position extracting means 73 determines whether or not the walk test is finished by the operator (step S3). If the human image position extracting unit 73 has not received a signal corresponding to the end operation of the walk test from the operation unit 40, the human image position extracting unit 73 returns the control to step S1 and repeats the processes of steps S1 to S3.

  On the other hand, when the human image position extracting unit 73 receives a signal corresponding to the end operation of the walk test from the operation unit 40, the area specifying unit 74 creates the appearance frequency distribution of the human image position stored in the storage unit 30 (step S4). When the area specifying unit 74 creates the appearance frequency distribution of the human image positions, the area specifying unit 74 obtains an edge position where the number of appearances rapidly decreases in the appearance frequency distribution of the created human image positions (step S5). Then, the area specifying unit 74 determines whether or not the edge position has been detected (step S6). When the area specifying unit 74 detects the edge position, the area specifying unit 74 estimates that the edge position or a position shifted from the edge position by a predetermined amount to the side where the human image position appears less frequently is the wall surface position. Then, the area specifying unit 74 sets an area having a size determined based on the viewing angle of the sensor unit 20, the angle of view of the imaging unit 10, and the like on the side where the appearance number of the human image position is large from the estimated wall surface position. It is specified as a region (sensor region) corresponding to the detection area (step S7).

  When the area specifying unit 74 specifies the sensor area, the area registration unit 75 generates sensor area information indicating the position of the sensor area and information outside the sensor area indicating the position of the area other than the sensor area, and registers the information in the storage unit 30. (Step S8).

  On the other hand, if the edge position cannot be detected in step S6, the area specifying unit 74 displays on the display unit 50 that the detection area of the sensor unit 20 has not been set correctly and notifies the operator (step S9). ).

Returning to FIG. 1, the abnormality monitoring unit 76 detects an intruding object that has entered the monitoring area based on the monitoring image received from the imaging unit 10 and the detection level information received from the sensor unit 20 in the alert mode. Then, when detecting the intruding object, the abnormality monitoring unit 76 transmits the intruding object detection signal and the monitoring image captured by the imaging unit 10 to the security device 3 via the communication unit 60.
In the present embodiment, the sensor area is an area where an intruding object is detected based on both the monitoring image and the detection level of the PIR sensor (hereinafter referred to as a combined area), and areas other than the sensor area are based only on the monitoring image. An example of a region for detecting an intruding object (hereinafter referred to as an image single area) will be described.

  Hereinafter, the processing in the alert mode controlled by the change area extraction unit 71 and the abnormality monitoring unit 76 will be described with reference to the flowchart shown in FIG. The process described below is executed at the timing when the change area extraction unit 71 receives a monitoring image from the imaging unit 10. In the alert mode, the sensor unit 20 regularly transmits detection level information to the abnormality monitoring unit 76. In the following, an example in which the abnormality monitoring unit 76 detects an intruder (person) as an intruding object will be described.

  The change area extraction unit 71 uses a difference process between frames or a background difference process from a plurality of monitoring images acquired continuously in time by the imaging unit 10, and a change area having a temporal change in luminance value. To extract. Then, the change area extraction unit 71 determines whether or not a change area having a predetermined size or more has been extracted (step S11). In addition, for example, this predetermined size shows a person that is determined by an experiment that actually images a person in a monitoring area so that the person can be distinguished from small animals such as cats, mice, and insects. The minimum size of the change area is set.

When the change area extraction means 71 extracts a change area of a predetermined size or larger, the abnormality monitoring means 76 calculates an evaluation value E indicating the degree of certainty that the change area is a person (step S12). In the present embodiment, the evaluation value E is calculated based on the feature quantity f ₁ representing the size of the change area and the feature quantity f ₂ representing the aspect ratio of the change area. The feature quantity f ₁ is the number of pixels included in the change area, and the feature quantity f ₂ is the ratio of the width and height of the circumscribed rectangle of the change area. When the anomaly monitoring means 76 calculates the feature quantities f ₁ and f ₂ , it approaches 1 as each feature quantity approaches a value determined based on the height and physique of a person, and approaches 0 as it moves away from the value. Normalized values F ₁ and F ₂ obtained by normalizing each feature amount are obtained. Then, the abnormality monitoring unit 76 sets a value obtained by multiplying the normalized values F ₁ and F ₂ as the evaluation value E.

Furthermore, the abnormality monitoring unit 76 uses a known tracking process for a change area extracted over a plurality of monitoring images continuously acquired by the imaging unit 10 and uses a known tracking process to associate the change area with the change area. The evaluation value E may be calculated by detecting an area and using information such as the moving direction and moving speed of the associated change area. In this case, for example, the abnormality monitoring unit 76 extracts, from the change area of interest extracted from the latest monitor image, the center of gravity position of the focus change area and the monitor image acquired one frame before. The distances from the center of gravity positions of all the changed areas are obtained. Then, the abnormality monitoring unit 76 sets the change area having the shortest distance between the center of gravity positions as the change area in which the same object as the object in the target change area is shown. Further, the abnormality monitoring unit 76 calculates the moving direction and moving speed of the target change area by calculating the difference between the center of gravity position of the target change area in the latest monitoring image and the corresponding change area in the previous monitoring image. . Then, the abnormality monitoring means 76 becomes closer to 1 as the movement direction obtained is closer to the movement direction obtained in the past with respect to the object reflected in the target change area (that is, the movement direction is straight ahead), A normalized value F ₃ that is closer to 0 as the distance from the determined moving direction is determined. The abnormality monitoring unit 76, the moving speed obtained becomes close to 1 closer to the walking speed of a human, obtaining a normalized value F ₄ to be close to zero as the distance from the walking speed of a human. Then, the abnormality monitoring unit 76 sets a value obtained by multiplying the normalized values F _{1 to} F ₄ as the evaluation value E.
The abnormality monitoring unit 76 is not limited to these examples, and the abnormality monitoring unit 76 evaluates evaluation values calculated by various known methods for calculating an evaluation value indicating the degree of certainty that the person is a predetermined area on the image. It may be used as the value E.

  After calculating the evaluation value E for the changed area, the abnormality monitoring unit 76 determines whether or not the changed area exists in the image single area (step S13). If the abnormality monitoring unit 76 determines that the change area exists in the image single area, the abnormality monitoring unit 76 determines whether or not the evaluation value E is greater than or equal to the reference value E0 (step S14). In this case, since an object shown in the change area does not exist in the detection area of the sensor unit 20, it is determined whether the change area is caused by an intruder with reference to the detection level P obtained by the sensor unit 20. Is not preferred. Therefore, the abnormality monitoring unit 76 determines whether or not the change area is caused by an intruder based only on the monitoring image. Therefore, the reference value E0 is set to, for example, a lower limit value of a value indicating that the change area is caused by an intruder, which is determined by an experiment for actually detecting the intruder based on the monitoring image.

  Then, when the evaluation value E is less than the reference value E0, the abnormality monitoring unit 76 determines that the change area is not caused by an intruder, and returns the control to step S11. On the other hand, when the evaluation value E is equal to or greater than the reference value E0, the abnormality monitoring unit 76 determines that the change area is caused by an intruder, that is, an intrusion abnormality has occurred (step S15). When the abnormality monitoring unit 76 determines that an intrusion abnormality has occurred, the latest monitoring image captured by the imaging unit 10 (hereinafter, the latest monitoring image when the abnormality monitoring unit 76 determines that an intrusion abnormality has occurred is referred to as an abnormal image). On the other hand, an image (hereinafter referred to as a composite image) displaying the detection area of the sensor unit 20 is generated. For example, the abnormality monitoring unit 76 reads the sensor area information from the storage unit 30 and combines a frame representing the combined area, that is, the range of the detection area, with the abnormal image to obtain a composite image. Then, the abnormality monitoring unit 76 transmits the intruding object detection signal and the composite image to the security device 3 via the communication unit 60 (Step S16).

  On the other hand, when the change area is present in the combined area, the abnormality monitoring unit 76 determines whether or not the detection level P received most recently from the sensor unit 20 is equal to or greater than a predetermined threshold Th (step S17). Note that the threshold value Th is set to a value indicating that an intruding object exists in the detection area based on experiments or experience. That is, when the detection level P is greater than or equal to the threshold Th, it can be determined that there is a high possibility that an intruder exists in the detection area based on the heat rays radiated from the detection area.

Therefore, when the detection level P is greater than or equal to the threshold value Th, the abnormality monitoring unit 76 determines whether or not the evaluation value E is greater than or equal to the reference value E1 (step S18). In this case, since there is a high possibility that the changed area is due to an intruder, the reference value E1 is set to a value lower than the reference value E0. That is, it becomes easier for the abnormality monitoring means 76 to determine that the change area is due to an intruder than when the change area is outside the combined area, and the intruder can be detected more reliably.
If the evaluation value E is less than the reference value E1, the abnormality monitoring unit 76 determines that the change area is not due to an intruder, and returns the control to step S11. On the other hand, when the evaluation value E is equal to or greater than the reference value E1, the abnormality monitoring unit 76 determines that the change area is caused by an intruder, that is, an intrusion abnormality has occurred (step S15), and generates a composite image. Then, the intruding object detection signal and the composite image are transmitted to the security device 3 via the communication unit 60 (step S16).

On the other hand, when it is determined that the detection level P is less than the threshold value Th, the abnormality monitoring unit 76 determines whether or not the evaluation value E is greater than or equal to the reference value E2 (step S19). In this case, since it is unlikely that the change area is due to an intruder, the reference value E2 is set to a value higher than the reference value E0. That is, the abnormality monitoring means 76 can suppress that it is difficult to determine that the change area is caused by an intruder and the erroneous detection of an object other than the intruder as an intruder than when the change area is outside the combined area. . This is because the presence of a person is not denied by the PIR sensor in the image alone area.
If the evaluation value E is less than the reference value E2, the abnormality monitoring unit 76 determines that the change area is not due to an intruder, and returns the control to step S11. On the other hand, when the evaluation value E is greater than or equal to the reference value E2, the abnormality monitoring unit 76 determines that the change area is caused by an intruder, that is, an intrusion abnormality has occurred (step S15), and generates a composite image. Then, the intruding object detection signal and the composite image are transmitted to the security device 3 via the communication unit 60 (step S16).

  The determination as to which region the change area belongs to in step S13 is that it is determined that the change area belongs to the combined area only when all or most of the changed area (for example, 70% or more) is included in the combined area. preferable. This is because, based on the detection principle of the PIR sensor, even if only a part of the human body enters the detection area, it is unlikely that a detection level P equal to or higher than the threshold value Th is obtained. Therefore, even if the changed area is partially included in the combined area, the intrusion determination is performed in accordance with the criterion of the image single area.

  The abnormality monitoring unit 76 determines whether or not the detection level P is equal to or higher than the threshold Th at the timing when the detection level information is received from the sensor unit 20, and even when the detection level P is equal to or higher than the threshold Th, step S11 is performed. The process of S19 may be performed. In that case, since it is considered that the change area extraction unit 71 has already extracted the change area for the latest monitoring image, the extraction process of the change area in step S11 is omitted, and the abnormality monitoring unit 76 determines the extracted change area. Steps S11 to S19 are performed again. Thereby, the intruding object detection device 1 can detect the intruder earlier, and can notify the security device 3 of the intrusion abnormality earlier.

  Further, the abnormality monitoring unit 76 does not synthesize a frame representing the combined area with the abnormal image, and transmits the abnormal image and the sensor area information together with the intruding object detection signal or the obstacle detection signal via the communication unit 60. May be sent to. In that case, the security device 3 generates a composite image from the abnormal image and the sensor area information, and transmits it to the monitoring center device 5. Alternatively, the security device 3 transfers the abnormal image and the sensor area information, and synthesizes and displays them in the monitoring center device 5.

  In addition, when the security device 3 receives the intruding object detection signal and the composite image, the security device 3 may transfer the received composite image to the operation display 4 and cause the operation display 4 to display the composite image.

  In addition, the abnormality monitoring unit 76 generates an image in which a frame representing a combined area is combined with the latest monitoring image captured by the imaging unit 10 in response to a request from the security device 3 or the operation indicator 4. 3 or the operation indicator 4 may be transmitted. Thereby, the security device 3 can transmit an image showing the combined area to the monitoring center device 6 in response to a request from the monitoring center device 6, and the operation indicator 4 displays the combined area according to the operation of the user. Displayed images.

  The processing to be performed separately for the sensor region and the region other than the sensor region is not limited to changing the reference value for comparison with the evaluation value E described above, but depends on the use, purpose, etc. for detecting the intruding object. May be determined as appropriate. For example, the abnormality monitoring unit 76 normalizes the evaluation value E and the detection level P in a range from 0 to 1 so that the evaluation value E and the detection level P are close to 0 as not to be an intruding object and close to 1 as the intruding object appears, and further the detection level P The threshold value Th ′ corresponding to the threshold value Th is obtained in the normalized range. Then, the abnormality monitoring means 76 adds the value obtained by normalizing the evaluation value E and the value obtained by normalizing the detection level P in the combined area, and adds the value obtained by normalizing the evaluation value E and the threshold value Th ′ in the image-only area. Then, it may be determined whether or not the change area is caused by an intruder by comparing with the same reference value.

  Further, the abnormality monitoring means 76 sets the reference value E0 to a sufficiently large value compared to the reference value E1 and the reference value E2 (that is, this value) in order to focus on the detection area (combination area) of the sensor unit 20. In this case, the relationship E1 <E2 << E0 is established), and it may be more difficult to determine that the intrusion abnormality is in the single image area than in the combined area. Thereby, the intruding object detection device 1 can suppress the detection omission of the intruder in the area to be monitored with priority, and can suppress erroneous detection due to planting, small animals, and the like in the other monitoring area.

  Alternatively, when the change area is in the sensor area, the abnormality monitoring unit 76 determines whether the change area is due to an intruding object based only on the detection level P from the sensor unit 20, and the change area is the sensor area. If the area is in a region other than that, it may be determined whether the changed region is due to an intruding object based only on the monitoring image. In that case, if the abnormality monitoring means 76 determines in step S17 shown in FIG. 4 that the detection level P is greater than or equal to the threshold Th, it determines that an intrusion abnormality has occurred and the detection level P is less than the threshold Th. If it is determined that there is no intrusion abnormality, it is determined. Thereby, the intruding object detection device 1 can suppress the false detection of the intruder in the detection area when planting or the like that reduces the detection accuracy of the intruder detection based on the monitoring image exists in the detection area. .

  Alternatively, the abnormality monitoring unit 76 may monitor the presence / absence of an intruder only in the detection area of the sensor unit 20, and may use an image only for the monitor to confirm in other areas. In that case, when the change area is in the sensor area, the abnormality monitoring means 76 determines whether or not the change area is caused by an intruding object based on the detection level P from the sensor unit 20 and the monitoring image. Is not in the sensor area, it is not determined whether the change area is due to an intruding object. In this case, the abnormality monitoring unit 76 may not extract the change region for the region other than the sensor region, and whether or not the evaluation value E is greater than or equal to the reference value E0 in step S14 shown in FIG. Regardless, the control may always be returned to step S11.

  As described above, in the intruding object detection device according to the embodiment of the present invention, the human image position extraction unit determines the presence or absence of an intruding object based on the detection level information acquired from the sensor unit, and determines that the intruding object exists. The position of the human image is extracted based on the change area on the monitoring image. Then, the area specifying unit detects a wall-side end portion of the movement range of the person on the monitoring image based on the appearance frequency distribution of the human image position, and specifies the detection area of the PIR sensor on the image. The abnormality monitoring means determines whether the change area is caused by an intruding object based on only the monitoring image when the change area exists outside the detection area, and detects at least when the change area exists within the detection area. Based on the level, it is determined whether or not the change area is caused by an intruding object. As a result, the intruding object detection device can distinguish the area corresponding to the detection area on the image from other areas with high accuracy, and can determine the presence or absence of the intruding object by appropriate processing in each area. The detection accuracy of the intruding object can be improved.

  The preferred embodiments of the present invention have been described above, but the present invention is not limited to these embodiments. For example, each time the human image position extracting unit acquires a monitoring image from the imaging unit, the human image position is extracted and stored in the storage unit regardless of the value of the detection level P, and the area specifying unit detects the edge position. After that, it may be determined whether or not the human image position exists further on the wall surface side than the detected edge position. Then, when the human image position is further on the wall surface side than the detected edge position, the area specifying means determines that the detection area of the sensor unit has not been set correctly and displays the fact on the display unit to notify the operator. To do. Thereby, the intruding object detection device can prevent an artificial setting error regarding the direction of the PIR sensor, and can improve the reliability of intrusion monitoring.

  In this embodiment, the example of obtaining the position of the detection area in the horizontal direction on the image has been described. However, a predetermined direction is determined according to the positional relationship between the imaging direction, the wall that is a restriction, and the detection area, such as the vertical direction and the oblique direction. It can be realized by replacing the direction. Moreover, you may use not only one direction but two directions.

Further, the abnormality monitoring means may change a criterion for determining whether or not the change area is an intruding object according to the moving path or the generation position of the change area that has occurred in the sensor area. In the intruding object detection device according to the present embodiment, it is possible to identify on which side the house exists on the basis of the sensor area on the image. For this reason, for example, when a person comes out of a house, or when a person looks out from a window of the house, the person is likely a resident of the house and is not an intruder. Can be configured not to report.
In that case, the abnormality monitoring means uses the tracking process for the change area extracted over a plurality of monitoring images continuously acquired by the imaging unit, and changes the change area that is associated with the same object. To detect. In the present embodiment, the change areas may be associated by the process described in step S12 of FIG. 5, or the change areas may be associated by using other known tracking processes.

And when the object suddenly appears at the position of the wall (near the wall) in the monitoring area, that is, when the newly extracted change area exists at the detection area boundary on the wall side which is the restricted object, It is recognized that it is caused by a person who has come out, and it is determined that the change area is not caused by an intruder. The abnormality monitoring unit also determines that the change area in the monitoring image of the subsequent frame corresponding to the change area determined not to be due to the intruder is not due to the intruder. In addition, the abnormality monitoring means corresponds to the change area determined not to be caused by the intruder, for the change area in the monitoring image of the subsequent frame, for example, after the predetermined period has elapsed, by the intruder as in the normal change area. You may make it determine whether it is a thing.
In addition, the abnormality monitoring means evaluates the change area corresponding to the object when the object moves from the wall side to the detection area or when the object suddenly appears near the boundary of the detection area on the wall side. The value E may be decreased so that it is difficult to determine that the change area is due to an intruding object. Furthermore, when the object suddenly appears near the boundary on the wall side of the detection area, the abnormality monitoring means decreases the evaluation value E as the position where the change region corresponding to the object appears closer to the wall surface position. Also good.

  Thereby, the intruding object detection device can determine the presence or absence of an intruding object based on the relationship between the moving path of the change region or the generation position and the wall surface position, and can further improve the detection accuracy of the intruding object.

In another embodiment, the area specifying unit 74 may estimate the end of the person's moving range using the detection level peak distribution instead of the human image position appearance frequency distribution. The entire configuration of the intruding object detection device in this case is the same as that shown in FIG. 1, and is realized by the human image position extraction means 73 and the area specifying means 74, as compared with the intruding object detection device according to the above embodiment. Only the function to be different is different. Therefore, only the differences from the intruding object detection device according to the above embodiment will be described below. The human image position extraction unit 73 determines whether or not the detection level P received from the sensor unit 20 is equal to or higher than a predetermined reference value Z2 during the walk test. When the detection level P is equal to or higher than the reference value Z2, the human image position extracting unit 73 extracts the center of gravity position of the change area extracted by the change area extraction unit 71 at that time as the human image position, and the detection level P at that time The data are stored in the storage unit 30 in association with each other.
The reference value Z2 is set to a value indicating that a person is present in the detection area, like the reference value Z1. Further, the reference value Z2 is preferably set to a value lower than the threshold value Th for determining an intruding object in the alert mode so that the change in the detection level P at each position can be observed.

  In this embodiment, after the end of the walk test, the area specifying unit 74 creates a peak distribution of detection levels at each horizontal position on the monitoring image based on the human image position and the detection level P stored in the storage unit 30. . For example, the area specifying unit 74 divides the monitoring image into a plurality of regions having a predetermined width in the horizontal direction, and for each divided region, the highest detection level P associated with the human image position indicating the position in the region. The value is extracted and set as a peak value. Alternatively, the area specifying unit 74 extracts, for each horizontal position on the monitoring image, the highest value among the detection levels P associated with the human image position recorded in a predetermined width area centered on the horizontal position, It may be a peak value corresponding to each horizontal position.

  FIG. 6 shows an example of a peak distribution of detection levels. In FIG. 6, a graph 600 represents a peak distribution of detection levels on the monitoring image, a horizontal axis represents a horizontal position on the monitoring image, and a vertical axis represents a peak value of the detection level at each position. In the graph 600, a range indicated by a solid line, that is, a range having a value equal to or higher than the threshold value Z2 indicates a region where a peak value is actually extracted. In the graph 600, the peak value of the detection level has a value greater than or equal to a certain value in a certain range on the left side of the coordinate X3, but rapidly decreases at the coordinate X3 and becomes zero in the range on the right side from the coordinate X3. That is, in this case, it can be estimated that the worker has not moved to a position corresponding to the position on the right side of the coordinate X3, and the wall surface position X4 exists in the vicinity of the coordinate X3.

Therefore, the area specifying unit 74 obtains an edge position where the peak value rapidly decreases with respect to the created peak distribution of the detection level. For example, the area specifying unit 74 sets a position where the peak value of the detection level at each horizontal position decreases by a predetermined threshold or more with respect to the peak value of the detection level at the adjacent position. Alternatively, the area specifying unit 74 generates an approximate curve of the peak value of the detection level using the least square method or the like, and the value by the approximate curve at each horizontal position is predetermined with respect to the value by the approximate curve at the adjacent position. A position that decreases more than a threshold value may be used as the edge position. Further, it may be added to the edge position determination condition that the peak value continues at 0 or a predetermined value or less at the horizontal position after the position where the peak value has suddenly decreased. Note that these predetermined threshold values can be determined appropriately by, for example, performing a walk test by experiment and measuring the peak value of the detection level.
Further, the area specifying unit 74 extracts an edge position based on the amount of decrease in the peak value of the detection level with respect to the position adjacent to the left side in the region on the right side of the center position on the monitoring image. In the left area, the edge position may be extracted based on the amount of decrease in the peak value of the detection level with respect to the position adjacent to the right side.

  Then, the area specifying unit 74 estimates that the wall position is a determined edge position or a position shifted by a predetermined amount from the edge position to the side where the peak value of the detection level is small, that is, the opposite side of the human movement range. Then, the area specifying unit 74 specifies a region having a predetermined size on the side where the peak value of the detection level is large from the estimated wall surface position, that is, the side where the person moves, as a region (sensor region) representing the detection area of the PIR sensor. . In the example shown in FIG. 6, the shaded portion 601 is specified as the sensor region.

  FIG. 7 shows a flowchart in the case where the area setting means 72 specifies the sensor region using the peak distribution of the detection level.

  In step S21 of the flowchart shown in FIG. 7, the human image position extracting unit 73 receives the detection level information from the sensor unit 20 as in step S1 shown in FIG. 4, and whether or not the detection level P is equal to or higher than the reference value Z2. Is determined (step S21). Then, when the detection level P becomes equal to or higher than the reference value Z2, the human image position extracting unit 73 extracts the center of gravity position of the changed area extracted by the changed area extracting unit 71 at that time as the human image position and stores it in the storage unit 30 (step S22). Thereafter, the human image position extracting means 73 determines whether or not the walk test has been finished by the operator (step S23).

When the human image position extracting unit 73 receives a signal corresponding to the end operation of the walk test from the operation unit 40, the area specifying unit 74 detects based on the human image position and the detection level P stored in the storage unit 30. A level peak distribution is created (step S24). After creating the peak distribution of the detection level, the area specifying unit 74 obtains an edge position where the peak value rapidly decreases with respect to the created peak distribution of the detection level (step S25).
Since the process of step S26-S29 is the same as the process of step S6-S9 shown in FIG. 4, description is abbreviate | omitted.

  In this embodiment, the area specifying unit detects the end of the moving range of the person on the captured image based on the peak value of the detection level detected by the sensor unit at each position on the captured image, and further detects the sensor region. Is identified. The intruding object detection device can reduce the influence of the measurement error of the thermal change in the detection area by obtaining the peak value of the detection level at each position on the captured image, and the range of the detection area on the captured image. Can be accurately identified. As a result, the intruding object detection device can improve the detection accuracy of the intruding object when determining the presence or absence of the intruding object by distinguishing the area corresponding to the detection area of the PIR sensor on the image from other areas. .

  As described above, those skilled in the art can make various modifications in accordance with the embodiment to be implemented within the scope of the present invention.

DESCRIPTION OF SYMBOLS 1, 2 Intruding object detection apparatus 3 Security apparatus 4 Operation indicator 5 Monitoring center 6 Local area network 7 Wide area communication network 10 Imaging part 20 Sensor part 30 Memory | storage part 40 Operation part 50 Display part 60 Communication part 70 Monitoring process part 71 Area setting Means 72 Change area extraction means 73 Human image position extraction means 74 Area identification means 75 Area registration means 76 Abnormality monitoring means 100 Security system

Claims (4)

A sensor unit that outputs a detection level corresponding to a heat ray radiated from the detection area, and a monitoring area that includes the restriction object and the detection area. An intruding object detection device having an imaging unit that acquires images at a predetermined cycle,
Change area extraction means for extracting a change area in the monitoring image;
Area setting means for executing a registration process for specifying a sensor area indicating the detection area on the monitoring image;
An abnormality monitoring means for executing an abnormality determination process for detecting an intruding object in the monitoring area based on the detection level and the monitoring image;
The area setting means includes
The position of the change area extracted from the monitoring image acquired by the imaging unit when the detection level output by the sensor unit is a value indicating that a person is present in the detection area. Human image position extracting means for extracting as:
Based on a plurality of the human image positions extracted by the human image position extracting means during a predetermined period, an end of the movement range of the person limited by the restricted object is detected on the monitoring image, and Area specifying means for determining the sensor region by defining a first end that is an end of the sensor region on the restricted object side based on the end, and
The intruding object detection device, wherein the abnormality monitoring means executes different abnormality determination processing depending on whether the change area exists outside the sensor area or the sensor area.   The area specifying means obtains the appearance frequency of the human image position extracted during the predetermined period for each position on the monitoring image, and sets a position where the appearance frequency rapidly decreases with respect to an adjacent position as the moving range. The intruding object detecting device according to claim 1, wherein the intruding object detecting device is an end of the intruding object. The human image position extracting means stores the human image position and the detection level in association with each other when the detection level is a value indicating that a person exists in the detection area,
2. The intrusion according to claim 1, wherein the area specifying unit sets a position at which the detection level stored in association with the human image position on the monitoring image rapidly decreases with respect to an adjacent position as an end of the moving range. Object detection device.   The abnormality monitoring unit determines that the change area is not caused by an intruding object when the newly extracted change area exists in the vicinity of the first end of the sensor area. The intruding object detection device according to any one of?

Images