JP6137234B2 - Detection device - Google Patents

Description

  The present invention relates to a detection device.

  2. Description of the Related Art Conventionally, an evacuation guidance system has been proposed that guides evacuation when a person in the office has a wireless transmission device and a disaster occurs in the office (see, for example, Patent Document 1).

JP 2005-338991 A

  However, in the conventional evacuation guidance system, a person in the office has to have a wireless transmission device. For this reason, for example, if the wireless transmission device is left on the desk, it may be erroneously detected that there is a person to be evacuated or the person who removed the wireless transmission device cannot be evacuated. There is a fear.

  Therefore, the present invention has been made in view of the above-described problems, and an object thereof is to provide a detection device capable of continuously detecting a subject with high accuracy.

The detection device of the present invention is different from the first imaging unit that captures an image including the subject from the first direction and the first imaging unit that captures an image including the subject from the first direction. The second imaging unit, the first imaging unit, and the second imaging unit so that the predetermined subject is included in a region where the respective imaging regions of the first imaging unit and the second imaging unit overlap. An adjustment unit that adjusts at least one of the imaging unit and a detection unit that detects a size of an image corresponding to the head of the subject by imaging at least one of the first imaging unit and the second imaging unit. And the relationship between the height of the subject and the size of the image corresponding to the subject's head, and personal identification information of the subject imaged by at least one of the first imaging unit and the second imaging unit And at least one of the first imaging unit and the second imaging unit. A determining unit posture information of the subject imaged on the basis of the magnitude of the change in the image corresponding to the head of the elephant's is the detection device comprising a.

In this case, the adjustment means can adjust an installation position or an optical system of at least one of the first imaging unit and the second imaging unit. Further, the adjusting means may use a human body as the predetermined subject. In this case, the adjusting means may use a human head as the predetermined subject.

In the detection apparatus of the present invention, the first at least one of the imaging unit the second imaging unit may be arranged stairs to the imaging positions. Further, the determination unit may determine the posture state of the subject from the change in the size of the image in consideration of the height position of the stairs where the subject exists.

  In the detection apparatus of the present invention, a tracking unit that tracks the target person may be provided by continuously using the imaging results of the imaging units in which a part of the imaging region overlaps. Moreover, when the subject is imaged by the first imaging unit, at least one of the start and end of imaging of the second imaging unit may be switched. The first direction may be a vertical direction, and the first imaging unit and the second imaging unit may capture the image from above the subject.

  The detection device of the present invention has an effect that the subject can be detected continuously with high accuracy.

It is a figure showing the office building where the guidance device concerning one embodiment was introduced. It is a block diagram which shows a guidance device. It is a block diagram which shows an image acquisition apparatus. It is a block diagram which shows the function implement | achieved when a program is performed in CPU. FIG. 5A is a diagram illustrating a hardware configuration of the host computer, and FIG. 5B is a functional block diagram of the host computer. It is a figure which shows the specific structure of an imaging part. FIG. 7A is a graph showing the relationship between the distance from the front focal point of the wide-angle lens system to the head of the person (subject) and the size of the image (head portion), and FIG. FIG. 8 is a graph obtained by converting the graph of FIG. 7A to a height from the floor. It is a graph which shows the change rate of the magnitude | size of an image. FIG. 9A and FIG. 9B are diagrams schematically showing changes in the size of the head image according to the posture of the subject. It is a figure which shows the change of the magnitude | size of the image of a subject's head imaged by an image pick-up element according to a subject's position. It is a figure which shows the data produced in an entrance process. It is a figure which shows typically the relationship between one division in an office building, and the imaging area of the imaging part of the image acquisition apparatus provided in the said division. FIGS. 13A and 13B are diagrams for explaining the tracking process in the case where four subjects (subjects A, B, C, and D) move within one section of FIG. (No. 1) FIGS. 14A to 14C illustrate tracking processing when four subjects (subjects A, B, C, and D) move within one section of FIG. (No. 2) Fig.15 (a)-FIG.15 (d) are the figures (the 1) for demonstrating the tracking method of the subject in the elevator EV. Fig.16 (a)-FIG.16 (c) are the figures (the 2) for demonstrating the tracking method of the subject in the elevator EV. It is a figure which shows the condition where the fire broke out in the office building. It is a flowchart which shows an evacuation guidance process.

  Hereinafter, an embodiment will be described in detail with reference to FIGS. FIG. 1 shows an example of an office building 200 in which a guidance device 100 (see FIG. 2) is introduced. In addition, the guidance device 100 can be introduced not only in an office building but also in a school, an apartment house, or the like.

  As shown in FIG. 1, the office building 200 is a 6-story building as an example, and a person can move between the floors (floors) using the stairs 24. Moreover, since the elevator EV is provided in the office building 200, it is possible to move between the floors by using the elevator EV. An elevator hall 41 is provided in the vicinity of the elevator EV on each floor. In addition, each floor has three sections (rooms) 43 as shown by being surrounded by broken lines in FIG. In the present embodiment, it is assumed that the entrance of the office building 200 is on the first floor.

  FIG. 1 shows a state in which the image acquisition device 2 included in the guidance device 100 is arranged on the ceiling of each floor. Here, one section 43 is a floor extending in the X-axis direction and the Y-axis direction in FIG. 1, and the section 43 has nine in total along the X-axis direction and three along the Y-direction. It is assumed that an image acquisition device 2 is provided. However, in FIG. 1, for convenience of illustration, a state in which three image acquisition devices 2 are installed for one section is illustrated. Note that the number of the image acquisition devices 2 can be appropriately determined based on the size of the section 43, the optical performance of the imaging unit included in the image acquisition device 2, and the like.

  FIG. 2 shows the guidance device 100 in a block diagram. As shown in FIG. 2, the guidance device 100 includes a plurality of image acquisition devices 2 provided on each floor and first to sixth CPUs (in a state where communication with the image acquisition devices 2 on each floor is possible). 3a-3f). Moreover, the guidance device 100 includes an image acquisition device 2 provided in the elevator EV, and an elevator CPU 3g to which the image acquisition device 2 in the elevator EV is connected. Furthermore, the guidance device 100 includes a host computer 4 to which the CPUs 3a to 3g are connected, and a reading device 20.

  The image acquisition apparatus 2 acquires an image including the head of a person (photographing (tracking) target person (hereinafter simply referred to as “target person”) within a predetermined range (in an imaging region) or On the other hand, an evacuation route is indicated. The detailed configuration of the image acquisition device 2 will be described later.

  The first to sixth CPUs (3a to 3f) and the elevator CPU 3g process information acquired by the image acquisition device 2 and control the operation of the image acquisition device 2. In this embodiment, one CPU is installed for each floor. However, the present invention is not limited to this, and a plurality of CPUs may be provided on each floor. In this case, information exchange between the CPUs may be performed by wired or wireless communication means. The detailed configuration of the CPUs 3a to 3g will be described later.

  When a disaster such as a fire occurs based on information output from the CPUs 3a to 3g, the host computer 4 notifies the person in the office of the occurrence of the disaster and guides evacuation.

  In FIG. 2, the guidance device when the office building 200 is 6 stories is illustrated. However, when the office building is 1 story, the CPUs 3 b to 3 g and the host computer 4 in FIG. 2 are omitted. May be. In this case, the function of the host computer 4 may be given to the first CPU 3a.

  For example, the reading device 20 is provided near the entrance of the office building 200 or in the vicinity of the reception. The reading device 20 has a card reader capable of reading information recorded on the IC card. The subject can read the personal information into the reading device 20 by holding the employee card (IC card) in which the personal information (personal identification information) is recorded in the internal memory over the reading device 20. If the target person is a guest, the guest information can be read into the reading device 20 by holding the guest certificate (IC card) passed at the reception over the reading device 20.

  Next, a specific configuration of the image acquisition device 2 will be described with reference to FIG. FIG. 3 shows a block diagram of the image acquisition device 2. As shown in FIG. 3, the image acquisition device 2 includes an imaging unit 8, an illuminometer 10, a pyroelectric sensor 12, a fire sensor 15, a microphone 13, a communication unit 9, and an LED (Light Emitting Diode). Diode) 11, flash memory 16, speaker 14, and processing unit 18. Further, the image acquisition device 2 includes a secondary battery 17. The image acquisition device 2 normally operates using a primary power source (for example, an office power source).

  The imaging unit 8 images the office building 200 and a person (target person) existing in the office building 200. A specific configuration of the imaging unit 8 will be described later.

  The illuminometer 10 measures the illuminance in the imaging area when the imaging unit 8 images the inside of the office. The measurement result obtained by the illuminometer 10 is input to the processing unit 18. The pyroelectric sensor 12 is a sensor for detecting whether a person is present in the imaging area of the imaging unit 8. As the pyroelectric sensor 12, an infrared sensor that detects infrared rays generated by a person, a pyroelectric sensor having a Fresnel lens, or the like can be used. The detection result by the pyroelectric sensor 12 is output to the CPUs (3a to 3g) via the processing unit 18 and the communication unit 9. In the present embodiment, the viewing angle (detection angle) of the pyroelectric sensor 12 is set according to the imaging region of the imaging unit 8 so that the imaging region of one imaging unit 8 can be covered by the single pyroelectric sensor 12. In addition, it is assumed that a detection visual field (detection distance) is set. However, the present invention is not limited to this, and a plurality of pyroelectric sensors 12 may be provided for each image acquisition device 2. Alternatively, one pyroelectric sensor 12 may be shared by a plurality of image acquisition devices 2.

  The fire sensor 15 is a sensor that detects smoke when a fire occurs, and includes, for example, a light emitting unit such as an LED and a light receiving unit such as a PD (Photo Diode). The fire sensor 15 detects a fire by receiving reflected light reflected by smoke particles out of the light emitted from the light emitting unit at the light receiving unit. The fire sensor 15 may be a sensor that detects the amount of heat when a fire occurs, instead of a sensor that detects smoke. In FIG. 3, one fire sensor 15 is provided for one image acquisition device 2. However, the present invention is not limited to this, and a plurality of fire sensors 15 may be provided for each image acquisition device 2. Good. Alternatively, a plurality of image acquisition devices 2 may share one fire sensor 15.

  The microphone 13 acquires the voice of a person (target person) and inputs it to the processing unit 18. For example, when the voice input from the microphone 13 is “help”, the processing unit 18 determines that rescue is necessary based on the voice.

  The communication unit 9 performs communication with the CPUs 3a to 3g and the like. Specifically, the communication unit 9 includes an antenna, a transmission unit, a reception unit, and the like. The CPU (3a to 3g) to which each image acquisition device 2 is directly connected is connected to the target person. In addition to transmitting information such as the status, communication with the host computer 4 and the communication unit 9 of another image acquisition device 2 is also performed.

  For example, when the office is partly turned off at night and the illuminance measured by the illuminometer 10 is low, the LED 11 is operated in accordance with an instruction from the processing unit 18 (for example, imaging of the imaging unit 8). Illuminate the area. The flash memory 16 is a storage medium that stores captured images and position information of subjects present in the office. The speaker 14 transmits information (such as disaster occurrence status information and evacuation route information) to the target person by sound or voice.

  The processing unit 18 performs processing based on the image captured by the imaging unit 8 and the detection results detected by the other units, and comprehensively controls the processing of each component in the image acquisition device 2.

  The secondary battery 17 is a backup power source that supplies power to the image acquisition device 2 when power is not supplied from the primary power source to the image acquisition device 2 due to the occurrence of a power failure or disaster. As the secondary battery 17, for example, a lithium ion battery can be used. However, the present invention is not limited to this, and as the secondary battery, solar power generation for offices or the like may be used.

  FIG. 4 is a block diagram illustrating functions realized by executing programs in the CPUs 3a to 3g. As illustrated in FIG. 4, the CPUs 3 a to 3 g function as a tracking unit 61, a posture determination unit 62, and a communication unit 63. In FIG. 4, the memory 64 connected to the tracking unit 61 and the posture determination unit 62 is also illustrated. After the subject enters from the entrance of the office building 200, the tracking unit 61 tracks the subject in a state where privacy is protected based on the imaging result of the imaging unit 8. The posture determination unit 62 determines the posture of the subject (upright, middle waist, lying down, etc.) in a state where the privacy of the subject is protected based on the imaging result of the imaging unit 8. The posture determination unit 62 also determines whether the subject is abnormal based on the change in posture. The communication unit 63 exchanges information with the communication unit 9 of the image acquisition device 2 and the host computer 4. For example, while the posture determination unit 62 determines that there is no abnormality in the subject, the communication unit 63 transmits a message to the host computer 4 that there is no abnormality in the subject. In this case, the image captured by the imaging unit 8 is not transmitted to the host computer 4. As a result, the privacy of the subject can be protected while the state where there is no abnormality continues. On the other hand, when an abnormality occurs in the subject, a message indicating that an abnormality has occurred is sent to the host computer 4, but in this case, an image may be sent to accurately convey the situation. it can.

  FIG. 5A shows the hardware configuration of the host computer 4. As shown in FIG. 5A, the host computer 4 includes a CPU 90, a ROM 92, a RAM 94, a storage unit (here, an HDD (Hard Disk Drive)) 96, an input / output unit 97, and the like. Each component of the host computer 4 is connected to a bus 98. In the host computer 4, the CPU 90 executes a program stored in the ROM 92 or the HDD 96, thereby realizing the functions of the respective units in FIG. The input / output unit 97 is connected to a sprinkler, an exhaust device, and a drive device for a protective wall (fire door) in the office. Sprinklers discharge water and extinguish it when a fire breaks out. Exhaust devices exhaust harmful gases (smoke, carbon monoxide, etc.) outside the office. Protective walls (fire doors) are doors for blocking the circulation of harmful gases, and are installed in hallways and stairs. The input / output unit 97 is connected to a communication device of a fire department or a security company by wire or wirelessly. Therefore, the host computer 4 can notify the fire department or the security company of the occurrence of a disaster via the input / output unit 97. 2 is connected to the input / output unit 97.

  FIG. 5B shows a functional block diagram of the host computer 4. As shown in FIG. 5B, the host computer 4 includes a setting unit 5, a calendar unit 6, a guidance control unit 51, a movement determination unit 53, a blocking unit 55, a route determination unit 57, an entrance processing unit 58, and a communication unit. 59 functions. The setting unit 5 performs various settings for the plurality of image acquisition devices 2. For example, the setting unit 5 sets an image acquisition interval of the image acquisition device 2. The calendar unit 6 has functions of acquiring calendar information such as year, month, day, hour, minute, and second and measuring time. The calendar information can be acquired from a CPU crystal oscillator included in the host computer 4 or a timing integrated circuit.

  The guidance control unit 51 controls the operation of each unit (for example, the LED 11) of the image acquisition device 2 in order to evacuate the target person based on the imaging result of the imaging unit 8 at the time of a disaster (such as when a fire occurs). The movement determination unit 53 determines whether or not the subject has moved as intended by the guidance control unit 51. The blocking unit 55 opens and closes the protective wall (fire door) in the office building 200 via the protective wall driving device based on the presence position of the subject obtained from the imaging result of the imaging unit 8. More specifically, when the size of the image of the subject's head is not detected from the imaging result of the imaging unit 8, the blocking unit 55 blocks the passage guided by the guidance control unit 51 with a protective wall. Moreover, the interruption | blocking part 55 interrupts | blocks the electric power supply to mechanical equipment at the time of a fire outbreak. The route determination unit 57 changes the evacuation route according to the situation in the office. The entrance processing unit 58 performs processing for setting a person who enters from the entrance of the office building 200 as a target person. The communication unit 59 transmits the processing results of the respective units to the communication unit 9 of the image acquisition device 2 via the communication units 63 of the CPUs 3a to 3g.

  Next, a specific configuration and the like of the imaging unit 8 included in the image acquisition device 2 will be described in detail based on FIG. The imaging unit 8 is mainly for imaging a person's head in an office building. Here, the ceiling height of each floor is assumed to be 2.6 m. That is, the imaging unit 8 images a human head or the like from a height of 2.6 m.

  As shown in FIG. 6, the imaging unit 8 includes a wide-angle lens system 32 having a three-group configuration, a low-pass filter 34, an imaging element 36 such as a CCD or a CMOS, and a circuit board 38 that drives and controls the imaging element. Have. Although not shown in FIG. 6, it is assumed that a mechanical shutter (not shown) is provided between the wide-angle lens system 32 and the low-pass filter 34.

  The wide-angle lens system 32 includes a first group 32a having two negative meniscus lenses, a second group 32b having a positive lens, a cemented lens, and an infrared cut filter, and a third group 32c having two cemented lenses. The diaphragm 33 is disposed between the second group 32b and the third group 32c. The wide-angle lens system 32 of this embodiment has a focal length of 6.188 mm and a maximum field angle of 80 °. The wide-angle lens system 32 is not limited to the three-group configuration. That is, for example, the number of lenses in each group, the lens configuration, the focal length, and the angle of view can be changed as appropriate.

  As an example, the image sensor 36 has a size of 23.7 mm × 15.9 mm and a pixel number of 4000 × 3000 (12 million pixels). That is, the size of one pixel is 5.3 μm. However, as the image sensor 36, an image sensor having a different size and the number of pixels from the above may be used.

  In the imaging unit 8 configured as described above, the light beam incident on the wide-angle lens system 32 enters the image sensor 36 via the low-pass filter 34, and the circuit board 38 converts the output of the image sensor 36 into a digital signal. Then, the processing unit 18 (see FIG. 3) including an ASIC (Application Specific Integrated Circuit) performs image processing such as white balance adjustment, sharpness adjustment, gamma correction, and gradation adjustment on the image signal converted into the digital signal. In addition, image compression such as JPEG is performed. Further, the processing unit 18 stores the JPEG compressed still image in the flash memory 16.

  Note that the imaging area of the imaging unit 8 overlaps with the imaging area of the imaging unit 8 of the adjacent image acquisition device 2 (see imaging areas P1 to P4 in FIG. 12). This point will be described in detail later.

  FIG. 7A is a graph showing the relationship between the distance from the front focal point of the wide-angle lens system 32 to the head of the person (subject) and the size of the image (head portion). FIG. 7B shows a graph obtained by converting the graph of FIG. 7A to the height from the floor.

  Here, as described above, when the focal length of the wide-angle lens system 32 is 6.188 mm and the diameter of the subject's head is 200 mm, from the front focal point of the wide-angle lens system 32 to the position of the subject's head. When the distance is 1000 mm (that is, when a person with a height of 1 m 60 cm stands upright), the diameter of the head of the subject imaged on the image sensor 36 of the imaging unit 8 is 1.238 mm. On the other hand, when the position of the subject's head is lowered by 300 mm and the distance from the front focal point of the wide-angle lens system 32 to the position of the subject's head is 1300 mm, an image is formed on the imaging device of the imaging unit 8. The diameter of the subject's head is 0.952 mm. That is, in this case, when the head height changes by 300 mm, the size (diameter) of the image changes by 0.286 mm (23.1%).

  Similarly, when the distance from the front focal point of the wide-angle lens system 32 to the position of the subject's head is 2000 mm (when the subject is in the middle waist), the subject's head that forms an image on the image sensor 36 of the imaging unit 8. Is 0.619 mm, and when the position of the subject's head is lowered by 300 mm, the size of the subject's head image formed on the image sensor of the imaging unit 8 is 0.538 mm. . That is, in this case, when the head height changes by 300 mm, the size (diameter) of the head image changes by 0.081 mm (13.1%). Thus, in the present embodiment, as the distance from the front focal point of the wide-angle lens system 32 to the subject's head increases, the change (change rate) in the size of the head image becomes smaller.

  FIG. 8 is a graph showing the change rate of the size of the head image. FIG. 8 shows the change rate of the image size when the position of the subject's head changes by 100 mm from the value shown on the horizontal axis. As can be seen from FIG. 8, when the distance from the front focal point of the wide-angle lens system 32 to the position of the subject's head is increased from 1000 mm to 100 mm, the change rate of the image size is as large as 9.1%. Even if the head size is the same, if the height difference is about 100 mm, a plurality of subjects can be easily identified based on the height difference. On the other hand, when the distance from the front focal point of the wide-angle lens system 32 to the position of the subject's head is away from 2000 mm to 100 mm, the change rate of the image size is 4.8%. In this case, although the rate of change of the image is smaller than when the distance from the front focal point of the wide-angle lens system 32 described above to the position of the subject's head is 1000 mm to 100 mm, the change in the posture of the same subject is reduced. If so, it can be easily identified.

  As described above, since the distance from the front focal point of the wide-angle lens system 32 to the subject can be detected from the size of the image of the subject's head by using the imaging result of the imaging unit 8 of the present embodiment, the CPU By using this detection result, the posture determination unit 62 (3a to 3g) can determine the posture of the subject (upright, lying in the middle, or lying down) and the change in posture. This point will be described in more detail based on FIG.

  FIG. 9A and FIG. 9B are diagrams schematically showing changes in the size of the head image according to the posture of the subject. As shown in FIG. 9B, when the imaging unit 8 is provided on the ceiling and the head of the subject is imaged, when the subject is standing upright like the subject on the left side of FIG. When the head is imaged large as shown in FIG. 9 (a) and is tilted like the subject on the right side of FIG. 9 (b), the head is imaged small as shown in FIG. 9 (a). In addition, when the subject is in the middle waist as in the center subject in FIG. 9B, the head image is smaller than when standing and larger than when lying down. Therefore, in the present embodiment, the posture determination unit 62 of the CPUs 3a to 3g detects the size of the image of the subject's head based on the image transmitted from the imaging unit 8, thereby changing the state of the subject. Can be determined. In this case, since the posture of the subject and the change in posture are discriminated from the image of the subject's head, privacy is protected compared to the case where discrimination using the subject's face or whole body is performed. Can do.

  7A, 7B, and 8 show graphs in the case where the subject is present at a position where the angle of view of the wide-angle lens system 32 is low (below the wide-angle lens system 32). ing. That is, when the subject is present at the peripheral field angle position of the wide-angle lens system 32, there is a risk of being affected by distortion according to the expected angle with the subject. This will be described in detail.

  FIG. 10 shows a change in the size of the image of the head of the subject imaged by the image sensor 36 in accordance with the position of the subject. It is assumed that the center of the image sensor 36 coincides with the optical axis center of the wide-angle lens system 32. In this case, even when the subject is standing upright, when the subject is standing directly below the imaging unit 8 and when standing away from the imaging unit 8, the imaging unit 8 is affected by distortion. The size of the image of the head imaged changes. Here, when the head is imaged at the position p1 in FIG. 10, the size of the image captured by the image sensor 36, the distance L1 from the center of the image sensor 36, and the center of the image sensor 36 are obtained from the imaging result. Can be obtained. In addition, when the head is imaged at the position P2 in FIG. 10, from the imaging result, the size of the image captured by the image sensor 36, the distance L2 from the center of the image sensor 36, and the center of the image sensor 36 are obtained. Can be obtained. The distances L1 and L2 are parameters representing the distance between the front focal point of the wide-angle lens system 32 and the subject's head. Further, the angles θ1 and θ2 from the center of the image sensor 36 are parameters representing the expected angle of the wide-angle lens system 32 with respect to the subject. In such a case, the posture determination unit 62 corrects the size of the captured image based on the distances L1 and L2 from the center of the image sensor 36 and the angles θ1 and θ2 from the center of the image sensor 36. In other words, when the subject is in the same posture, the size of the image captured at the position p1 of the image sensor 36 is corrected so as to be substantially equal to the size of the image captured at the position p2. By doing in this way, in this embodiment, regardless of the positional relationship between the imaging unit 8 and the subject (distance to the subject or the prospective angle with the subject), the posture of the subject can be accurately detected. Can do. It is assumed that parameters (correction table) used for this correction are stored in the memory 64.

  By the way, the imaging unit 8 also has a function as a normal camera in addition to the above-described function of acquiring the image of the subject's head. That is, for example, by adding a zoom lens to the wide-angle lens system 32, for example, when the subject falls on the floor, the subject's eyes and mouth can be continuously imaged. In this case, the processing unit 18 uses a pattern matching method or the like to determine whether the eyes are open or whether the eyeball is moving, whether the mouth is moving, or breathing from the movement of the mouth. It can be determined whether or not.

  Further, the imaging unit 8 can apply contrast AF that extracts a high-frequency component of a signal from the image sensor 36, detects a lens position where the high-frequency component is maximized, and performs focus detection. In this case, an image can be acquired in a focused state by adjusting a part of the wide-angle lens system 32.

  Here, the imaging interval by the imaging unit 8 is set by the setting unit 5 of the host computer 4. The setting unit 5 can change the shooting frequency (frame rate) in a time zone where there is a high possibility that many people are in the office and in other time zones. For example, when the setting unit 5 determines from the calendar information acquired from the calendar unit 6 that it is currently a time zone in which there is a high possibility that there are many people in the office (for example, from 9:00 am to 6:00 pm). In this case, a still image is captured once per second (32,400 images / day). If it is determined that the time is other than that, a still image is captured once every five seconds (6480 images / day). ) Can be set.

  In the present embodiment, since the image acquisition device 2 includes the pyroelectric sensor 12, the processing unit 18 controls the imaging timing of the imaging unit 8 based on the output of the pyroelectric sensor 12. More specifically, the processing unit 18 does not perform imaging of the imaging unit 8 that is determined that no person is present in the imaging region based on the output of the pyroelectric sensor 12. On the basis of the output, the imaging of the imaging unit 8 determined to have a person in the imaging area is started. As a result, it is possible to reduce the power consumed by the imaging unit 8 and save the recording area of the flash memory 16. From the viewpoint of saving the recording area of the flash memory 16, the captured still image is temporarily stored in the flash memory 16, and then, for example, the captured image data for each day is transferred to the host computer 4, and then the flash memory You may make it erase from 16.

  In the present embodiment, since the image acquisition device 2 includes the illuminometer 10, the processing unit 18 uses, for example, the illuminometer 10 to determine the imaging time and ISO sensitivity when the imaging unit 8 captures a still image. Is determined based on the detected illuminance. In this case, for example, when the office is dark, the ISO sensitivity may be increased and the imaging time may be increased. If the office lighting cannot be used at all due to a power failure or the like, the CPU (3a to 3g) or the host computer 4 switches the power supply of the image acquisition device 2 from the primary power supply to the secondary battery 17, and then performs processing. The unit 18 may illuminate the office with the LED 11 and perform imaging using the imaging unit 8 in that state. In this case, the processing unit 18 may confirm the position of the subject by the pyroelectric sensor 12 described above, and may adjust the illumination direction of the LED 11 so that illumination light can be emitted to the position. In such a case, it is possible to take an image while the target person is efficiently illuminated. The LED 11 may be provided on the ceiling, may be provided at an arbitrary position in the office, or another light source may be used instead of the LED.

  Furthermore, the still image when the subject is imaged only requires an image that can detect the abnormality of the subject, and is different from the still image that is normally viewed. For this reason, at the time of shooting in a dark situation, the infrared cut filter disposed in the wide-angle lens system 32 may be retracted by a retracting mechanism (not shown) to secure the light quantity.

  Next, processing by the guidance device 100 configured as described above will be described.

(Normal (non-disaster) processing)
First, the processing at the normal time, that is, at the time of non-disaster will be described.

(1) Entrance Process Hereinafter, the entrance process when the subject enters the office building 200 will be described. The admission processing unit 58 of the host computer 4 controls the image acquisition device 2 (imaging unit 8) provided at the entrance of the office building 200 (on the first floor as described above), and is in an upright state near the entrance. Take a picture of the head of the visitors. In addition, the host computer 4 uses the reading device 20 to acquire the personal information of the visitors from the employee ID card (ID card) held by the visitors. As a result, the personal information (personal ID, etc.) of the visitors and the size of the head image when standing upright (the size of the head image when standing upright represents the size of the head itself in relation to height. It is also possible to acquire data in which the height is associated (see FIG. 11). The height may be associated with personal information in advance, or may be detected using another imaging device, a distance sensor, or the like when photographing the head. If the latter method is used, the height can be acquired even if the visitor is a guest. As another imaging apparatus, an imaging apparatus that photographs the subject from the horizontal direction can be employed. If the distance between the other imaging device and the subject is always taken in a constant state, the height of the subject can be calculated from the imaging result. A laser range finder or the like can be used as the distance sensor.

  Further, the entrance processing unit 58 stores the head image of the subject in the internal memory, and registers the file name of the image file as the head image in the data of FIG.

  In the present embodiment, since the height of the attendant in an upright state is associated with the size of the head image (see FIG. 11), the tracking unit 61 of the CPUs 3a to 3g selects the attendant. As information for tracking, a graph as shown in FIG. 7B can be prepared for each visitor. Note that the graph of FIG. 7B is a graph used for tracking the second visitor from the top of FIG. 11 (the visitor whose ID is 1786033). For other visitors, a graph suitable for the relationship between the height and the size of the head image is prepared.

(2) Tracking process of the target person on the first floor Next, the tracking process of the target person on the first floor after the target person (the attendee who has performed the entrance process) enters the office building 200 will be described. Here, under the control of the host computer 4, the tracking unit 61 of the CPU 3a tracks the subject.

FIG. 12 is a diagram schematically showing, as an example, the relationship between one section 43 in the office building 200 and the imaging area of the imaging unit 8 of the image acquisition device 2 provided in the section 43. . In FIG. 1, the case where nine imaging units 8 are provided in one section 43 has been described, but here, for convenience of explanation, four imaging units 8 (imaging regions P <b> 1) in one section 43. , P2, P3, and P4 are shown). One section is assumed to be 256 m ² (16 m × 16 m). Further, each of the imaging areas P1 to P4 is assumed to be a circular area, and is overlapped (overlapped) with adjacent imaging areas in the X direction and the Y direction. In FIG. 12, for convenience of explanation, a divided portion obtained by dividing one section into four (corresponding to the imaging regions P1 to P4) is shown as divided portions A1 to A4. In this case, assuming that the angle of view of the wide-angle lens system 32 is 80 °, the focal length is 6.188 mm, the height of the ceiling is 2.6 m, and the height of the subject is 1.6 m, centering directly below the wide-angle lens system 32. The inside of a circle having a radius of 5.67 m (about 100 m ² ) is an imaging region. That is, since the divided portions A1 to A4 are 64 m ² , the divided portions A1 to A4 can be included in the imaging regions P1 to P4 of each imaging unit 8 and a part of the imaging region of each imaging unit 8 is included. It is possible to overlap.

  FIG. 12 shows the concept of overlapping (overlap) of the imaging areas P1 to P4 as viewed from the object side. The imaging areas P1 to P4 are areas where light is incident on the wide-angle lens system 32. Not all of the light incident on the light enters the rectangular image sensor 36. For this reason, in the present embodiment, the imaging unit 8 may be installed in the office so that the imaging regions P1 to P4 of the plurality of adjacent imaging elements 36 overlap (overlap). Specifically, an adjustment unit (for example, a long hole, a large adjustment hole, or a shift optical system that adjusts the imaging position) that adjusts the attachment of the imaging unit 8 is provided, and images captured by the respective imaging elements 36. It is sufficient to adjust the overlap (overlap) while confirming with the eye and determine the mounting position of each imaging unit 8. For example, when the divided portion A1 shown in FIG. 12 and the imaging area of the imaging device 36 match, the images captured by the respective imaging units 8 do not overlap each other and are exactly matched. . However, considering the degree of freedom in attaching each of the plurality of imaging units 8 and the case where the installation height differs depending on the ceiling beam or the like, the imaging regions P1 to P4 of the plurality of imaging devices 36 overlap (over) as described above. It is preferable to wrap).

  The overlap amount can be set based on the size of the person's head. In this case, for example, if the outer periphery of the head is 60 cm, a circle having a diameter of about 20 cm may be included in the overlapping region. In addition, under the setting that only a part of the head needs to be included in the overlapping region, for example, a circle having a diameter of about 10 cm may be included. If the overlapping amount is set to this level, the adjustment when the imaging unit 8 is attached to the ceiling is facilitated. In some cases, the imaging regions of the plurality of imaging units 8 can be overlapped without adjustment.

  Hereinafter, a tracking process when four subjects (subjects A, B, C, and D) move within one section 43 in FIG. 12 will be described with reference to FIGS. 13 and 14. Since each subject is continuously tracked after entering the office building 200, it is assumed that the CPU 3a knows who the head of each subject is.

  FIG. 13A shows the state at time T1. FIGS. 13B to 14C show states after time T1 (time T2 to T5).

  At time T1, the subject person C exists in the divided portion A1, and the subjects A and B exist in the divided portion A3. In this case, the imaging unit 8 having the imaging region P1 images the head of the subject C, and the imaging unit 8 having the imaging region P3 images the heads of the subjects A and B.

  Next, at time T2, the imaging unit 8 having the imaging region P1 images the heads of the subjects B and C, and the imaging unit 8 having the imaging region P3 images the heads of the subjects A and B.

  In this case, the tracking unit 61 of the CPU 3a moves the subjects A and C from the imaging results of the imaging units 8 at times T1 and T2 in the horizontal direction of FIG. It recognizes that it is moving up and down in b). The reason why the subject B is captured by the two imaging units 8 at time T2 is that the subject B exists in a portion where the imaging regions of the two imaging units 8 overlap.

  Next, at time T3, the imaging unit 8 having the imaging region P1 images the heads of the subjects B and C, and the imaging unit 8 having the imaging region P2 images the head of the subject C and has the imaging region P3. The imaging unit 8 images the subject A's head, and the imaging unit 8 having the imaging region P4 images the subjects A and D's head.

  In this case, the tracking unit 61 recognizes that the subject A is at the boundary between the divided portion A3 and the divided portion A4 (moving from the divided portion A3 to the divided portion A4) at time T3, and the subject B Recognizes that the user is in the divided part A1, recognizes that the subject C is at the boundary between the divided part A1 and the divided part A2 (moving from the divided part A1 to A2), and the subject D is the divided part. Recognize being at A4.

  Similarly, at time T4, the tracking unit 61 indicates that the subject A is in the divided portion A4, the subject B is in the divided portion A1, the subject C is in the divided portion A2, and the subject D is between the divided portions A2 and A4. Recognize In addition, at time T5, the tracking unit 61 recognizes that the target person A is between the split part A4, the target person B is between the split part A1, the target person C is between the split part A2, and the target person D is between the split part A2. .

  In the present embodiment, as described above, since a part of the imaging regions of the plurality of imaging units 8 is overlapped, the tracking unit 61 can recognize the position and moving direction of the subject. Thus, in the present embodiment, the tracking unit 61 can continuously track each target person on the first floor with high accuracy. The tracking unit 61 can also grasp the number of subjects on the first floor of the office building 200 from the number of subjects being tracked.

  When the size of the head image changes in such continuous tracking of the subject, the posture determination unit 62 determines the size of the head image and the data shown in FIG. Data) associated with the person) can be used to determine the posture of the subject (falling, lying in the middle, standing upright, etc.).

  The processing unit 18 may perform calibration (calibration) between the imaging units 8 with a part of the imaging region using the imaging results of the imaging units 8 with a part of the imaging region overlapping. . That is, when the same subject's head is imaged in two imaging units in which a part of the imaging region overlaps, the processing unit 18 is different even if the size of the head image is different. The imaging results of at least one of the imaging units are calibrated so that they match. Thereby, it is possible to suppress the influence of the imaging error between the imaging units.

(3) Target Person Tracking Process in Elevator EV Next, a target person tracking method in the elevator EV will be described with reference to FIGS. 15 and 16.

  When the subject moves between floors by the elevator EV, the subject is tracked by the tracking unit 61 of the CPU 3g and the tracking unit 61 of the CPUs 3a to 3f.

  Here, as an example, as shown in FIG. 15A, five people (subjects A to E) ride on the elevator EV on the first floor, and the elevator EV moves on the second floor as shown in FIG. 15B. It is assumed that the elevator EV stops on the third floor as shown in FIG. 15 (c). Then, as shown in FIG. 15 (d), on the third floor, three people (subjects A to C) get off from the elevator EV, and two people (subjects F and G) get on, as shown in FIG. 16 (a). As shown, it is assumed that the elevator EV passes through the fourth floor as it is, and two persons (subjects D and E) get off the elevator EV on the fifth floor as shown in FIG. Then, as shown in FIG. 16 (c), it is assumed that the remaining two persons (subjects F and G) get off the elevator on the sixth floor.

  In such a case, the imaging unit 8 provided on the ceiling of the elevator EV images the heads of five people riding from the first floor, and the imaging results of the tracking unit 61 of the first CPU 3a on the first floor and the elevator CPU 3g. Output to the tracking unit 61. In addition, the output to the tracking part 61 of 1st CPU3a is performed via the communication part 63 of CPU3g for elevators. In this case, the tracking unit 61 of the first CPU 3a uses the subjects A to E based on the results of imaging the subject who has been waiting for the elevator EV in the elevator hall 41 on the first floor and the imaging results of the imaging unit 8 in the elevator EV. Recognizes that it has started to move on the elevator EV from the first floor.

  In addition to this, instead of the above, the tracking unit 61 of the first CPU 3a on the first floor has the heads of the subjects A to E in the imaging results captured by the imaging unit 8 present in the elevator hall on the first floor. It may be made to recognize that subjects A to E started moving on the elevator EV based on the fact that it is no longer possible to acquire the image.

  The tracking unit 61 of the first CPU 3a on the first floor transmits the result recognized as described above to the tracking unit 61 of the elevator CPU 3g. The elevator CPU 3g can recognize from the recognition result of the tracking unit 61 of the first CPU 3a on the first floor that the person on the elevator is the target person A to E.

  15D, after the elevator stops on the third floor, the tracking unit 61 of the third CPU 3c on the third floor and the tracking unit 61 of the elevator CPU 3g perform imaging results before and after opening and closing of the elevator EV. Pass each other. That is, while the tracking unit 61 of the third CPU 3c can no longer recognize the head images of the subjects F and G, immediately after that, the tracking unit 61 of the elevator CPU 3g outputs the images of the heads of the subjects F and G. As a result, the tracking unit 61 of the elevator CPU 3g recognizes that the target persons F and G have entered the elevator from the third floor. On the other hand, the tracking unit 61 of the third CPU 3c can newly recognize the images of the heads of the subjects A to C, but at this point of time, it cannot recognize who the recognized head is. . Accordingly, when the tracking unit 61 of the elevator CPU 3g cannot capture the images of the heads of the subjects A to C in the elevator EV, the tracking unit of the third CPU 3c on the third floor captures the images of the heads of the subjects A to C. By outputting to 61, the third CPU 3c performs pattern matching processing between the images of the heads of the subjects A to C input from the elevator CPU 3g and the images captured by the imaging unit 8 provided in the elevator hall on the third floor. Thus, it is recognized that the person who got down to the third floor from the elevator EV is the target person A to C. The tracking unit 61 of the elevator CPU 3g may output an image before the elevator stops on the third floor and the door is opened to the tracking unit 61 of the third CPU 3c. Thereby, the tracking unit 61 of the third CPU 3c can recognize in a shorter time that the person who has descended from the elevator EV to the third floor is the target person A to C.

  16B and 16C, information is exchanged between the CPU tracking unit 61 of each floor and the CPU tracking unit 61 of the elevator EV in the same manner as described above. By performing the processing as described above, it is possible to track a person who moves with the elevator EV.

  In addition, in the image acquisition apparatus 2 provided in the elevator EV and the elevator hall 41, it is good also as monitoring raising / lowering of a subject's elevator by pattern matching with the head image acquired at the time of entrance.

  In the above description, the target person who gets on and off the elevator EV is recognized from the head image. However, the present invention is not limited to this, and the target person may be recognized from the size of the head image. . This recognition method is particularly effective when there is a large difference in the size of the head image of the subject riding on the elevator EV when standing upright.

(4) Tracking process of the subject on the stairs For the tracking on the stairs 24, basically the same method as the tracking process of the target person on the first floor is used. However, if there is a step in the height direction (Z direction) like a staircase, the head of the imaged image will depend on where the subject is on the staircase, even though it is standing upright. The size of the image changes. For this reason, simply acquiring an image of the subject's head and recognizing the posture of the subject based on the image may result in the loss of the subject's posture or inability to correctly recognize the posture or posture change of the subject. There is a risk.

  Therefore, in the present embodiment, the step information of the stairs and the imaging position of the imaging element of the imaging unit 8 are stored in correspondence with each other in the flash memory 16 of the image acquisition device 2 provided on the stairs. For example, the flash memory 16 stores data in which the position in the XY plane of each stage captured by the imaging unit 8 is associated with the height at the position. Specifically, the flash memory 16 associates the height from the 0th step of the staircase (for example, 170 mm) with the position in the XY plane of the first step staircase in the XY plane of the second step staircase. The height from the second step of the stairs (for example, 340 mm) is associated with the position in the table, and data such as.

  Accordingly, the processing unit 18 of each CPU can detect the posture of the target person in consideration of the position (XY position) of the stairs where the target person exists and the height of the stairs. The posture can be detected with high accuracy.

  Since the staircase is narrow but has a step, the wide-angle lens system 32 of the imaging unit 8 provided on the staircase may be changed to a wide-angle lens system for staircases. As the basic configuration of the wide-angle lens system in this case, the same configuration (three-group configuration) and the same number of lenses as described above can be adopted, but the difference in height from the first floor to the second floor (about 3500 mm to 4000 mm). ) To change the focal length (for example, 9 mm). By doing in this way, it is possible to track the target person with high accuracy. Note that the number of the image acquisition devices 2 for the stairs can be appropriately selected according to the height of the stairs, the position of the landing, and the like.

(5) Target person tracking process on the second to sixth floors On the second to sixth floors, the same process as the tracking process on the first floor described above is performed. In this case, the target person is continuously tracked using the result of the tracking process at the elevator or the stairs.

(6) Abnormality determination based on the posture of the subject When the processing unit 18 normally tracks the subject, the state in which the subject falls down continues for a predetermined time (determined by the posture judgment unit 62). Then, it is determined that an abnormality has occurred in the subject. In addition, the process part 18 can acquire the time when the subject fell into the state which fell down from the attitude | position determination part 62 of CPU3a-3g, and the time when the said state is continuing.

  However, in the office, the business may be performed for a long time with the subject seated. In order to cope with this, for example, the size of the image of the head when the subject is sitting on the chair is stored in the flash memory 16, and the processing unit 18 determines that the size of the image of the head is sitting on the chair. When it is maintained for a long time at the size of, it is determined that there is no abnormality because the sitting state in the chair continues. In this case, the flash memory 16 may store the position information (XY position information) where the desk is arranged and the image pickup position of the image pickup device 36 in association with each other. This makes it possible to perform abnormality determination with higher accuracy.

(7) Imaging timing control of the imaging unit 8 based on the presence / absence of the target person As described above, the imaging timing control using the pyroelectric sensor 12 can be performed, but the adjacent imaging areas are overlapped. Thus, the imaging timing can also be controlled.

  For example, when there are few people in the office, the processing unit 18 of each of the CPUs 3a to 3g triggers that a person's head is imaged in an overlapping area of a certain imaging unit 8 and corresponds to the overlapping area. Imaging by another imaging unit 8 is started. By doing in this way, when there is no person in the imaging area, imaging can be stopped (power supply to the imaging unit 8 is stopped), so that power consumption can be reduced. Further, the pyroelectric sensor 12 can be omitted.

  In addition, since overlapping a part of the imaging areas of the plurality of imaging units 8 is also adopted in the plurality of imaging units 8 provided on the stairs, the imaging units 8 provided near the stairs and the stairs are provided. The above-described control may be performed with the imaging unit 8 that has been provided. Thereby, power saving can be achieved also in the imaging unit 8 provided on the stairs. In this case, if the staircase is an emergency staircase, it is not so often used in normal times, so power saving is particularly effective.

(Processing when a disaster (fire) occurs)
When a disaster (fire) occurs, in addition to the normal processing as described above, the following processing is performed. In the present embodiment, as shown in FIG. 17, it is assumed that a fire has occurred in a section on the second floor + Y end of the office building 200. Here, only the first to third floors of the office building 200 are shown for the sake of simplicity.

  In this case, the guidance control unit 51 of the host computer 4 comprehensively controls all the image acquisition devices 2 via the CPUs 3a to 3g so that all persons in the office building 200 are evacuated outside the office building 200. Execute the process to guide as much as possible.

  Specifically, the guidance control unit 51 executes processing according to the flowchart of FIG.

  In the process of FIG. 18, in step S10, the guidance control unit 51 of the host computer 4 waits until a disaster is confirmed. In this case, as shown in FIG. 17, if a fire has occurred on the second floor of the office building 200, the fire sensor 15 of the image acquisition device 2 located in the vicinity of the fire detects the fire. Is output to the second CPU 2 b and the host computer 4 via the communication unit 9. Thereby, the guidance control part 51 can confirm that the fire broke out in the + Y end section on the second floor. Thus, when a fire is confirmed, judgment of step S10 is affirmed and it transfers to step S12. The guidance control unit 51 acquires the fire occurrence time from the calendar unit 6, and the calendar unit 6 starts measuring time from the fire occurrence time. This is because, for example, 10 minutes after the occurrence of a fire is set as the evacuation end time when performing evacuation guidance, as will be described later. In addition, the guidance control part 51 is good also as shortening the imaging interval of the imaging part 8 rather than the normal time, when a disaster is confirmed. In addition, when a disaster is confirmed, the guidance control unit 51 may switch the imaging by the imaging unit 8 from still image capturing to moving image capturing.

  Next, in step S12, the guidance control unit 51 operates the sprinkler near the fire occurrence location via the input / output unit 97 to extinguish the fire, and fires are generated by the speakers 14 of all the image acquisition devices 2. Announce in the office.

  Next, in step S14, the guidance control unit 51 confirms the number of people in the office building 200 acquired by the CPUs (3a to 3f) and the elevator CPU 3g on each floor. In addition, each CPU3a-3g can acquire the number of persons from the head image which each image acquisition apparatus 2 acquired. At this time, the guidance control unit 51 determines whether or not the number of persons confirmed as described above and the number of persons who have entered the office building 200 by subtracting the number of persons who have exited as a result of tracking are the same. Check it. In this case, there is no particular problem when the number of people is the same, but when the number of people is not the same, the guidance control unit 51 recognizes the number and performs the tracking process by the tracking unit 61 described above. Identify people who are not tracked. In addition, as a reason which cannot be tracked, the case where the image acquisition apparatus 2 close | similar to a fire spot is broken, or the head cannot be imaged with smoke etc. are assumed, for example.

  Next, in step S16, the blocking unit 55 stops the driving of the machinery that has been determined to stop driving in advance when a disaster occurs. In this case, the blocking unit 55 stops the machinery in an appropriate state. For example, the blocking unit 55 stops the elevator EV at a nearby floor while avoiding the second floor where the fire is occurring, opens the door, and stops subsequent driving.

  Next, in step S18, the guidance control unit 51 determines whether evacuation is necessary. Here, the guidance control unit 51 determines whether or not evacuation is necessary based on an image sent from the image acquisition device 2 near the fire site or whether the image acquisition device 2 can be driven by the influence of a fire. Can do. If the judgment here is negative, that is, if the fire extinguishes by the sprinkler and the fire subsides, and the evacuation is no longer necessary, the process proceeds to step S20. On the other hand, if the determination is affirmed, that is, if evacuation is necessary, the process proceeds to step S24.

  When it transfers to step S20, the guidance control part 51 announces that there is no need for evacuation by the speaker 14. FIG. Moreover, the interruption | blocking part 55 restarts the drive of the stopped machinery in step S22, and complete | finishes all the processes of FIG.

  On the other hand, when the process proceeds to step S24, the route determination unit 57 determines the evacuation route. In this case, the route determination unit 57 performs evacuation guidance in order from the target person located near the fire occurrence site. In this embodiment, as shown in FIG. 17, the place where the fire occurred is given first priority (the place indicated by the circled number 1 in FIG. 17), and the section adjacent to the upper and lower sections of the place where the fire occurred next is given second priority. (Places with circled number 2), the section on the floor above the fire location and the center section on the first floor have the third priority (the place with the circled number 3), and the other sections have the fourth priority (circle) Number 4). The reason why the central section on the first floor is given the third priority is to avoid the concentration of people at the exit on the first floor when the people on the upper floor evacuate. Note that the priority setting is not limited to the above setting method, and various setting methods can be employed. For example, the route determination unit 57 obtains a distance from a reference position (for example, a fire occurrence location) based on the position of the subject (detected by the processing unit 18), and determines a guidance method based on the distance. It is good as well. Even in this way, it is possible to guide the subject appropriately. In addition to the priority order, the route determination unit 57 does not use the stairs close to the fire occurrence site, for example, and uses the other stairs so that the person does not concentrate on one stairs. Can be determined.

  Further, the route determination unit 57 determines whether or not the target person can pass through the passage provided in the office building 200 based on the imaging result of one of the imaging units 8, and based on the determination result. The guidance route may be determined. In this case, the path determination unit 57 compares the sample image (reference image) captured in advance in the normal time with the image at the time of the disaster (that is, the image captured after capturing the sample image) (pattern matching). ) And it is good also as judging whether a subject can pass through a passage from the size of those differences. Note that the case where the difference is large includes, for example, the case where the document shelf is collapsed. Therefore, by determining as described above, it is possible to provide an appropriate evacuation route to the target person.

  The route determination unit 57 may only display the pattern matching result between the sample image and the image at the time of the disaster on the display device 91 together with the image at the time of the disaster. In this case, as a result of pattern matching, the display priority may be increased as the degree of divergence between the sample image and the image at the time of the disaster increases, and the imaging result may be displayed based on the priority. The person who has seen the content displayed on the display device 91 confirms the image at the time of the disaster with reference to the display priority, and changes the evacuation route when it is determined that the target person cannot pass through the passage (appropriate If an appropriate evacuation route is input to the host computer 4), the subject can evacuate along an appropriate evacuation route.

  Next, in step S26, evacuation guidance is started. Specifically, the guidance control unit 51 calls for evacuation to the outside using a part other than the stairs in the vicinity of the fire occurrence site from each of the speakers 14 of the image acquisition device 2 in the area of the circled numeral 1 in FIG. At this time, the guidance control unit 51 may visually indicate the evacuation route by sequentially blinking the LEDs 11 of each image acquisition device 2 along the evacuation route. And the guidance control part 51 performs evacuation guidance one by one with respect to the person who exists in the area | region of the round numbers 2-4.

  Next, in step S <b> 28, the guidance control unit 51 reports a fire to the fire department via the input / output unit 97. Then, in the next step S30, the movement determination unit 53 confirms the evacuation situation using the imaging result of the image acquisition device 2 provided on each staircase. Here, for example, even in a situation where the image acquisition device 2 near the fire occurrence site cannot be imaged, the movement determination unit 53 can estimate the evacuation situation from the imaging result of the image acquisition device 2. As described above, since the target person is always tracked in the office building 200, the target person who can be tracked can grasp the evacuation situation, but the target is not tracked. This is because the image acquisition device 2 suddenly starts imaging for a person. That is, in the present embodiment, among subjects who are to be imaged (scheduled to be imaged) in any of the imaging units 8, targets that are not imaged by any imaging unit at a certain time (first time) By specifying the person, the target person who is not imaged by the imaging unit at the certain time point (first time point) and is imaged at the certain time point (second time point) is identified at the first time point. It is possible to determine that the subject has been selected.

  Next, in step S32, the guidance control unit 51 confirms whether there is a person who needs to be rescued. In this case, when the size and position of the image of the head being imaged are almost unchanged in the determination result of the posture determination unit 62, this means that the target person is a person who cannot move. Therefore, the guidance control unit 51 uses the determination result of the posture determination unit 62 to determine whether there is such a head image. When rescue is necessary, a voice such as “help” may be input from the microphone 13 of the image acquisition device 2. Therefore, the guidance control unit 51 may determine that there is a person who needs to be rescued when such a voice is input. Furthermore, since the imaging unit 8 can continuously capture the subject's eyes and mouth when the subject is lying on the floor, the guidance control unit 51 can continuously A person who cannot move (for example, a fainting person) may be determined based on the captured image. If the determination in step S32 is negative, the process proceeds to step S36. If the determination is positive, the process proceeds to step S34.

  When the process proceeds to step S34, the guidance control unit 51 performs a process such that the rescue operation is executed. Specifically, the guidance control unit 51 displays the layout of the office building 200 on the display device 91 of FIG. 2 and blinks an area where there is a person who needs help. And the guidance control part 51 notifies the area | region (section) where the person who needs help exists to the fire department via the input-output part 97. FIG. In addition, the guidance control unit 51 displays the time measured (elapsed time) started after confirming the disaster in step S10 on the display device 91, and also notifies the fire department of the time measured (elapsed time). After step S34, the process returns to step S32.

  On the other hand, when the process proceeds to step S <b> 36, the movement determination unit 53 determines whether all have evacuated. Specifically, the movement determination unit 53 determines how many subjects have moved out of the office building 200 based on the imaging result of the image acquisition device 2 provided near the entrance (exit) on the first floor (end evacuation). Check). In addition, the movement determination part 53 is good also as confirming the number of persons remaining in the office building 200 from the imaging result of the image acquisition apparatus 2 provided in places other than the entrance (exit) vicinity of the 1st floor. In this case, the movement determination unit 53 determines the number of heads lost and the number of people who have performed entrance processing when all the head images are lost in the images captured by all the image acquisition devices 2 in the office building 200. Can be determined that everyone in the office building 200 has evacuated. If the determination in step S36 is negative, the process returns to step S32.

  Thereafter, the processes and determinations in steps S32 and S36 (or S34) are repeated until the determination in step S36 is affirmed.

  In step S36, the number of lost heads does not match the number of people who have performed entrance processing even though all the images of the heads have been lost in the images captured by all the image acquisition devices 2. It can happen. In such a case, there is a high possibility that there is a person who has not evacuated in the area where the image acquisition device 2 that cannot capture images due to a fire is installed. Therefore, in such a case, the guidance control unit 51 causes the display device 91 to blink and display the section where the image acquisition device 2 that cannot be imaged is present in the next step S34, and to notify the fire department of the same content. To do.

  In the process of FIG. 18, since the evacuation status is confirmed in step S30, the blocking unit 55 blocks the protective wall so as to close the passage where the subject does not exist based on the confirmation result. Can be.

  As described above in detail, according to the present embodiment, the imaging unit (first imaging unit) 8 that captures an image including the subject from above (+ Z direction) and the subject from above (+ Z direction). An imaging unit (second imaging unit) 8 different from the first imaging unit that captures an image including the image, and a processing unit 18 that detects size information of the subject's head from the image captured by the imaging unit 8; Among the plurality of imaging units 8, a part of the imaging region of the imaging unit 8 (first imaging unit) and a part of the imaging region of the imaging unit 8 (second imaging unit) Duplicate. Thereby, even when a subject moves between imaging regions by using both imaging units having overlapping imaging regions, the subject can be detected continuously with high accuracy.

  Further, according to the present embodiment, since the overlapping amount is an amount based on the size of the head, the size information of the head image is continuously displayed even when the subject moves across the imaging region. Can be detected with high accuracy.

  Further, according to the present embodiment, since the posture determination unit 62 determines the posture state of the subject based on the detection result by the processing unit 18, the posture state of the subject is continuously determined with high accuracy. Can do.

  In this case, the posture determination unit 62 acquires the height of the subject, the size of the subject's head, and personal information as data, and determines the subject from the size of the detected image of the subject's head. Therefore, it is possible to determine the posture of the subject in consideration of the height of each subject and the size of the head image.

  In the present embodiment, at least one of the plurality of imaging units is arranged at a position where the staircase can be imaged, and the posture determination unit 62 considers the height position of the staircase where the target person exists. Thus, the posture state of the subject can be determined from the size of the image of the subject's head. Thereby, the posture of the subject can be determined without being affected by the height of the stairs.

  Moreover, in this embodiment, since the target person is tracked by continuously using the imaging result of the imaging unit where a part of the imaging area overlaps, the target person who moves across the imaging area is continued, It is possible to track with high accuracy.

  Moreover, in this embodiment, when a subject is imaged in one imaging unit among the plurality of imaging units 8, imaging of an imaging unit having an imaging region overlapping with a part of the imaging region of the imaging unit is performed. Since at least one of the start and end is switched, it is possible to reduce the power consumption of the imaging unit and to start imaging of the imaging unit at an appropriate timing.

  Further, since the processing unit 18 can also calibrate the imaging unit with a part of the imaging region using the imaging result of each imaging unit with the part of the imaging region overlapping, an imaging error between the imaging units is possible. Can be suppressed.

  In the present embodiment, since the posture information of the subject is acquired using the size of the subject's head image taken from the vertical direction (+ Z direction), the head image taken from another direction is obtained. Compared with the case of using, the position of the subject's head in the vertical direction (that is, posture information) can be detected with high accuracy.

  In the above-described embodiment, the case where the imaging unit 8 is provided on the ceiling and the head of the subject is imaged from above has been described. However, it is not restricted to this, You may provide the imaging part 8 in the position which can image | photograph a subject from the side. By imaging the subject from the side, it is possible to easily discriminate movements of the eyes, mouth, nose, and the like. Furthermore, if it is determined whether the subject is normal using both the still image of the imaging unit 8 provided on the ceiling and the still image obtained by imaging the subject from the side, it is possible to make a highly accurate determination. Become.

  In the above embodiment, the image sensor 36 of the imaging unit 8 may be a CMOS, and an electronic shutter (rolling shutter) may be employed instead of the mechanical shutter. In this case, it is possible to suppress the generation of sound (shutter sound) during imaging. Thereby, the influence on the subject's smooth business performance by imaging can be suppressed.

  In the above-described embodiment, the case where the imaging unit 8 captures a still image has been described. However, the present invention is not limited thereto, and a moving image may be captured. In this case, moving images may be taken continuously, or short moving images of about 3 to 5 seconds may be taken intermittently. When the imaging unit 8 captures a moving image, the processing unit 18 may perform MPEG processing on the image signal and record it in the flash memory 16.

  In the embodiment described above, the guidance control unit 51 changes the imaging interval of the imaging unit 8 when a disaster occurs (step S10). However, the present invention is not limited to this. For example, when the posture determination unit 62 determines that an abnormality has occurred in the subject, the imaging interval of the imaging unit 8 may be changed.

  In the above-described embodiment, the case where the position and orientation of the target person are determined based on the imaging result of the imaging unit 8 has been described. However, the present invention is not limited to this. It is good also as detecting by the apparatus (laser range finder etc.) which detects a subject's position by irradiating light and receiving the light reflected by the subject.

  In the above embodiment, the imaging region of the imaging unit 8 provided in the elevator EV and the imaging region of the imaging unit 8 provided in the elevator hall 41 may be overlapped. Thereby, it is possible to track the movement of the subject between the elevator EV and the elevator hall 41 by the same method as in FIGS. 13 and 14.

  In the above embodiment, the case where the height is measured by imaging the subject from the side is exemplified, but the present invention is not limited to this. For example, an office structure (for example, a door) and a subject are imaged simultaneously, and the height of the subject is estimated using known structure dimensions (for example, the height of the door is 2 m and the width is 1 m). You may make it do. In this case, the dimensions of the structure may be recorded in the flash memory 16.

  In addition, although description is abbreviate | omitted in the said embodiment, in the private space (for example, a toilet etc.), the tracking by the imaging part 8 will be stopped. Even in such a case, tracking can be resumed by re-identifying the person who has come out of the private space from the height, the size of the head image, the head image, and the like. Further, when the subject is hidden under a desk or the like when an earthquake occurs, the same processing as described above can be performed.

  In the above-described embodiment, it is possible to appropriately confirm whether or not the target person being tracked is erroneously recognized. This confirmation can be performed in the office building 200 using information (such as a personal ID) when entering and leaving the room using the employee ID, and information such as logging into a personal computer.

  In the above embodiment, the blocking unit 55 controls the driving of the protective wall (fire door). However, the present invention is not limited to this, and the driving of the exhaust equipment such as an exhaust duct is controlled. Also good.

  In the above embodiment, the processing unit 18 detects the size of the head as the size information of the target person, but the present invention is not limited to this. For example, the processing unit 18 may detect a shoulder width or the like as the size information of the subject. In this case, the amount of overlapping of the imaging regions can be an amount based on the size related to the shoulder width or the like.

  Moreover, although the said embodiment demonstrated the case where an object person was guide | induced to the exterior of the office building 200 from the 1st floor entrance of the office building 200 at the time of evacuation, it is not restricted to this. For example, in the case of a high-rise building, it may be evacuated to the roof (such as a heliport). Further, the evacuation place and evacuation method may be changed as appropriate according to the type of disaster. Furthermore, the evacuation site may be different for each floor of the office building 200.

  In addition, although the said embodiment demonstrated the case where the guidance apparatus of this invention was used for evacuation at the time of a disaster, it is not restricted to this. For example, it can also be used when all members of the office building 200 are guided to a meeting place or a hall.

  The above-described embodiment is an example of a preferred embodiment of the present invention. However, the present invention is not limited to this, and various modifications can be made without departing from the scope of the present invention.

8 Imaging Unit 18 Processing Unit 51 Guidance Control Unit 53 Movement Determination Unit 57 Route Determination Unit 61 Tracking Unit 62 Posture Determination Unit

Claims (9)

A first imaging unit that captures an image including the subject from a first direction;
A second imaging unit different from the first imaging unit that captures an image including the subject from the first direction;
At least one of the first imaging unit and the second imaging unit is included so that a predetermined subject is included in a region where the respective imaging regions of the first imaging unit and the second imaging unit overlap. Adjusting means for adjusting;
A detection unit that detects a size of an image corresponding to the subject's head by imaging at least one of the first imaging unit and the second imaging unit;
A relationship between the height of the subject and the size of the image corresponding to the head of the subject, and personal identification information of the subject imaged at least one of the first imaging unit and the second imaging unit; Determination for determining posture information of a subject imaged based on a change in size of an image corresponding to the head of the subject imaged by at least one of the first imaging unit and the second imaging unit And
Ru detection device equipped with a.   The detection device according to claim 1, wherein the adjustment unit adjusts an installation position or an optical system of at least one of the first imaging unit and the second imaging unit.   The detection apparatus according to claim 1, wherein the adjustment unit uses a human body as the predetermined subject.   The detection apparatus according to claim 3, wherein the adjustment unit uses a human head as the predetermined subject. Said first of said second at least one of the imaging unit and the imaging unit, the detection apparatus according to claim 1, characterized in that it is arranged stairs to the imaging positions. The determination unit, in consideration of the height position of the staircase the subject exists, from the change in the size of the image, to claim 5, characterized in that to determine the posture state of the subject The detection device described. The imaging result of the imaging unit in which a part of the imaging regions overlap with continuous use it, the detection of any one of claim 1 to 6, characterized in that it comprises a tracking unit for tracking the subject apparatus. When the subject by the first imaging unit is captured, to any one of claims 1 to 7, characterized in that switch the start and at least one end of the imaging of the second image pickup unit The detection device described. The first direction is a vertical direction;
The detection apparatus according to any one of claims 1 to 8 , wherein the first imaging unit and the second imaging unit capture the image from above the subject.

Images