JP3495146B2 - Monitoring device using image analysis - Google Patents

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a surveillance device using image analysis, particularly an intruder into a controlled area such as an airport or a nuclear facility,
The present invention relates to a system configuration of a device that constantly monitors an intruder day and night.

[0002]

2. Description of the Related Art Conventionally, in a controlled area such as an airport, a nuclear facility, a special building or facility, monitoring is performed by a monitoring device using an infrared sensor or the like. For example, a plurality of pole-shaped (columnar) main bodies each having an infrared sensor are arranged at a monitoring place around a building / facility, and each of these has a predetermined height (for example, 10
Infrared sensors are placed on the pole-shaped body to send and receive infrared rays. According to this, the presence of an intruder or an intruder can be detected by detecting the light-shielding state in which the transmission and reception of infrared rays is temporarily interrupted.

On the other hand, a surveillance camera is arranged above to photograph a wide area including a plurality of surveillance areas between the pole-shaped bodies, and the surveillance camera sequentially scans a predetermined surveillance area while scanning at a constant speed, for example. I'm shooting and shooting important parts. Then, when an abnormality is detected by the infrared sensor of the pole-shaped body, the surveillance camera is automatically moved to the area between the pole-shaped bodies. It will be possible to monitor with a monitor such as a center.

[0004]

By the way, in the above-described conventional monitoring device, after the abnormality is detected by the infrared sensor, until the surveillance camera moves and an abnormal image is displayed,
It took several seconds to several tens of seconds, and there was a problem that abnormal images could not be observed well. That is, it is preferable that the surveillance camera being scanned accidentally includes an abnormal area, but the longer the distance from the abnormal area is, the longer it takes to move the surveillance camera. Therefore, during the movement of the camera, the intruder and the intruder are separated from the place, and the intruder or the like cannot be reliably captured by the surveillance camera.

Further, since the infrared sensor determines whether or not a thin infrared beam is blocked, it is difficult for one infrared sensor to reliably detect a large intruder or the like. Therefore, when a bird, an animal, or the like enters, erroneous detection is likely to occur. Therefore, the applicant has a built-in camera in the pole-shaped body to detect an intruder by comparing and analyzing an actual captured image, improve the detection accuracy, and instantaneously display an image of an abnormal area. A monitoring device that can be used is proposed (Japanese Patent Application No. 5-341508).
issue).

However, although there is no problem when an infrared camera is used as the above-mentioned camera, in order to reduce the cost, when an ordinary (visible light) camera is used, it is possible to detect an intruder image at night. There is a problem that you cannot do it.

Therefore, the applicant has proposed to detect an intruder by photographing a light beam spot and determining the light blocking state of the light beam spot on an image, as described later. In this case, for example, even if an intruder dressed in various colors is required to maintain high detection accuracy, the pole-shaped main body as a monitoring device may be slightly tilted due to an earthquake or the like. Sometimes, it is preferable to maintain a good detection state.

The present invention has been made in view of the above problems, and an object thereof is to enable day and night monitoring even with an ordinary camera, and to be highly effective against intruders dressed in various colors. It is an object of the present invention to provide a monitoring device that uses image analysis that maintains the detection accuracy and that can ensure a good detection state even when a slight change occurs in the pole-shaped body.

[0009]

In order to achieve the above object, a monitoring device utilizing image analysis according to the invention described in claim 1 projects a plurality of light beams into a monitoring area at predetermined intervals. For the light emitting unit and the main body around this light emitting unit.
A black second area, a plurality of light spots projected from the light projecting section, and a surveillance camera (normal camera or near infrared camera) for photographing the surveillance area including the second area ; An image comparison circuit for sequentially comparing the image of the light beam spot and the image of the second area of the black color obtained by the surveillance camera, and a light shielding (dimming) state of the light beam based on the output of the image comparison circuit and the second Area change (brightening)
It includes a determination circuit for determining condition, the determination in the determination circuit
It is characterized in that an abnormal state in the surveillance area is detected from the shielded state of the light rays and the change state of the second area, and the video signal from the surveillance camera can be output to the monitor when the abnormality occurs. For example, in the second area above,
For example, a black material is placed, and the brightness of this black area changes.
Will be detected.

[0010] invention of the second aspect, wherein the positional deviation amount determining circuit for determining the positional deviation amount of the upper Symbol beam spots provided by the positional deviation amount determining circuit is determined that the spot position is shifted by a predetermined amount At this time, at least the spot area which is the object of the image comparison is automatically reset. Here, it is preferable to reset the second area at the same time. The invention according to claim 3 is characterized in that a frame memory for storing one frame image when the above-mentioned abnormal state is detected is provided, and the image in this frame memory can be output to a monitor and held. To do.

[0011]

According to the structure described in the first aspect, the light blocking state of a plurality of light beam (for example, infrared) spots captured by the camera is determined by image comparison, and based on this, an abnormal state in which an intruder exists Is detected. In this case, since the plurality of light beam spots are arranged at a predetermined interval, only an object having a certain size is detected, and a bird or a small animal and an intruder can be well distinguished. Then, at the same time that an alarm is issued in the event of an abnormality, the image of the intruder or the like taken by the surveillance camera is displayed on the monitor of the center, for example, and is recorded in a recording device such as a VTR at night. The above abnormal state can be determined when the infrared sensor is separately arranged in the main body and the light blocking state of the infrared sensor and the light blocking state of the light beam spot are determined.

Also, the circumference of the projected light spot area is
The image change is also determined for the black second region of the circle . For example, when the color of the clothes of the intruder is similar to the above-mentioned light ray color, or when the clothes are whitish in the daytime, it is difficult to determine the light blocking state of the light spot.
In such a case, in the second area, the white clothes are arranged in the black area, so that the luminance change is detected. Therefore, it is possible to cover the state that is difficult to judge by the light spot and improve the detection accuracy.

According to the second aspect of the invention, even if the position of the light spot from the light projecting portion is displaced due to an earthquake, ground movement, or the like, a spot area for automatically following and comparing images is provided. It will be reset to the shifted position.

According to the third aspect of the invention, since one frame image at that time is stored in the frame memory at the time of abnormality, the image at the moment when the intruder is detected can be supplied to the external monitor. It enables quick confirmation and response.

[0015]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS FIG. 1 shows the structure of a composite type monitoring device using image analysis according to the embodiment, and FIG. 2 shows the arrangement and structure of the device. The abnormal state is detected by the determination by the infrared sensor and the determination by the image analysis. First, as shown in FIG. 2, the pole-shaped main bodies 10A and 10B, together with the other main bodies 10C, have a predetermined distance D (about 100 m at maximum).
The light emitting units 11 and 12 for emitting and receiving infrared rays are arranged outwardly as infrared sensors below the respective pole-shaped main bodies 10.

In addition, the other main body 10 can be provided by infrared rays whose rays are narrowed down (otherwise, red light, blue light, etc. may be used).
The light projecting portions 13 to 16 for forming a light beam spot on the side are also arranged outward. Then, a normal (visible light) camera (for example, having a 10x zoom function) 18 is attached as a monitoring camera to a position facing the two light projecting portions 15 and 16 of the other main body 10, while the other main body 10 is attached. A similar camera 19 is arranged at a position facing the two light projecting units 13 and 14 of 10. Therefore, the area of the distance D is
Lower infrared sensor (11, 12) and two cameras 1
8 and 19 will be monitored. Furthermore, the main body 1
The 0A and 10B are provided with two strobes 20 and 21 and a control circuit 22 that controls the entire apparatus.

FIG. 1 shows the configuration of the main circuit of the above embodiment, and the figure shows the circuit for an infrared sensor and one surveillance camera. In FIG.
Is composed of a light emitting signal generation circuit 24 and an infrared LED [light emitting diode (may be a blue light emitting diode or an infrared light emitting diode)] 25. In the light emitting signal generation circuit 24, a periodic signal is obtained by dividing the oscillation signal. By adding
A pulse-coded light projection signal is formed, and an infrared pulse is output from the infrared LED 25 based on this light projection signal. In addition, the light receiving unit 12 includes a light receiving element (photodiode) 26, an amplifier circuit 27, a waveform shaping circuit 28, and a signal recognition circuit 29. The infrared light received by the light receiving element 26 is amplified, waveform shaped, and transmitted. The signal recognition circuit 29 determines whether or not it is a signal.

On the other hand, the control circuit 22 includes the camera 1
Image processing circuit 3 for inputting video signal from 8 (19)
1, a latch circuit 32, a memory (writeable / readable) 33 for storing image information of, for example, one field, and an image data comparison circuit 34. The image processing circuit 31 amplifies and clamps the video signal. Then, it is converted into a digital signal. Then, the latch circuit 32 latches the data of the spot area in the previously obtained image data, the memory 33 stores the current image data of one field, and the image data comparison circuit 34 latches the latched image data. And the data of the spot area read from the memory 33 are compared with each other, and a change between images having a time difference is detected. That is, if an intruder or the like passes through, the light beam is blocked. Therefore, the light blocking state is determined by comparing the image data of the light beam spot area, for example, the brightness data.

Then, a judgment circuit 35 consisting of a logical operation circuit is provided at the subsequent stage of the image data comparison circuit 34. In this judgment circuit 35, the output of the signal recognition circuit 29 and the output of the image data comparison circuit 34 are provided. It is determined from the abnormal state. For example, when the signal recognition circuit 29 cannot recognize the transmitted infrared light pulse, it outputs a high level signal, and the image data comparison circuit 34 also performs image comparison analysis to determine that the spot area described later is in a light-shielded state. If a high-level signal is output, the decision circuit 3
No. 5 is determined to be abnormal when both outputs are at high level according to the AND condition.

A speaker 37 is connected via a sound output circuit 36 to the latter stage of the judging circuit 35, and an external output circuit 38 is provided in the judging circuit 35. Strobes 20, 21
A lighting signal is sent to and a warning signal is also supplied to the center. Therefore, when an abnormality occurs, the speaker 37 issues an alarm and at the same time, a signal indicating the abnormality is supplied to the center and the strobes 20 and 21 are turned on. The lighting of the strobes 20 and 21 may be controlled so as to be limited to nighttime.

Further, a frame memory 40 (IC memory) for storing one frame of image data is arranged at the subsequent stage of the memory 33, and when the judgment circuit 35 judges an abnormal state, the memory 33 is stored. Enter the latest field data currently stored in
The image data of the frame is stored. As described above, since the strobes 20 and 21 emit light at night, the image illuminated by the strobe is stored in the frame memory 40. The image data in the frame memory 40 is
The external output 38 is preferentially supplied to the center or the like via the switch circuit 41.

On the other hand, in the subsequent stage of the image processing circuit 31,
A luminance modulation circuit 42 for inputting the clamped video (analog) signal is connected, and the video signal output from the luminance modulation circuit 42 is supplied to the external output circuit 38 via the buffer amplifier 43 and the switch circuit 41. Therefore, in the embodiment, the image taken by the camera 18 can be supplied to the monitor or recording device at the center at any time, and at the time of abnormality, the image at that moment is preferentially supplied from the frame memory 40 via the switch circuit 41. Will be.

Further, in the embodiment, a spot position setting circuit 45 and a light beam spot positional deviation amount judging circuit 46 are provided. The spot position setting circuit 45 specifies an image of the light beam spot and judges the positional deviation amount. The circuit 46 calculates the positional deviation amount of the spot, and when the positional deviation amount within the predetermined range occurs, the spot position is reset.

That is, FIG. 3 shows the structure of the light projecting portion and the spot position on the screen. As shown in FIG. 3A, the light projecting portions 15 (13) and 16 (14) are , Lens W
The black plates 48A and 48B are disposed around the lens W, that is, in the upper part in the embodiment. Then,
As shown in FIG. 3B, the light spots S1 and S2 projected from the light projecting units 15 and 16 are reflected by the camera 18 (19).
Caught in. Therefore, the spot position setting circuit 45
Then, in the field memory 33 corresponding to the screen 70, a square (circular shape or the like, which is determined by the coordinate values (x1, y1) and (x2, y2) in a diagonal positional relationship, corresponding to the light beam spot S1 may be used. ) Area L1
A square area L2 determined by the coordinate values (x3, y3) and (x4, y4) is set corresponding to the light spot S2, and the spots are specified by specifying the addresses of these areas L1 and L2. The images in the areas L1 and L2 will be read out.

In the embodiment, the light beam spot S1
And above the black plate 48 on the upper side (or lower side) of S2.
The second areas M1 and M2 are set corresponding to the positions of A and 48B, and the second area M1 is determined by the coordinate values (x5, y5) and (x6, y6) which are in the diagonal positional relationship. Square area and second area M2 have similar coordinate values (x7, y7)
And a square area determined by (x8, y8). Then, if the addresses of these areas M1 and M2 are designated, the images of the second areas M1 and M2 are read.

Therefore, the images of the spot regions L1 and L2 and the second regions M1 and M2 are compared by the image data comparison circuit 34, and the light shielding state in the monitoring area is determined by the comparison. It That is, the intruder can be identified when the spot areas L1 and L2 are both shielded (dimmed) or when the change (increased) is seen in both the second areas M1 and M2. It is considered to be in a light-shielded state or a changed state. According to the second areas M1 and M2, it is possible to detect an intruder that is difficult to distinguish in the spot areas L1 and L2.

FIG. 4 shows the positional deviation state of the light beam spot S and the area for the positional deviation judgment. In the positional deviation amount judgment circuit 46, the judgment area C1 is set as a position deviation area which can be followed, and the judgment is made. The area C2 is an abnormal area that does not follow. Further, in this circuit 46, the positional deviation of the light beam spot S is determined by image comparison (for example, change in luminance amount) of the spot region L, and within a short time, for example, ±.
A deviation within 15% is considered to be a temporary shaking, and is not judged as a positional deviation.

Therefore, when the light beam spot S, which was at the center position at the time of initial setting, moves to the position of SP, for example, in the region C1, it is judged to be a permanent change such as an earthquake or ground movement. The spot area L is automatically reset at the position of SP. At the same time, the second area M is also automatically reset to the upper portion of the spot area L. However, when the light beam spot S moves to the position of SQ in the area C2, it is determined that the light spot S is in an abnormal state such as collapse, and a signal indicating this is supplied to the center.

The embodiment has the above construction, and its operation will be described with reference to the operation flowchart of FIG. First, at the time of initial setting, in step F1, the spot position setting circuit 45 sets the spot regions L1 and L2 in accordance with the light beam spots S1 and S2 of FIG. Next, when the monitoring operation is started in step F2, it is determined in the next step F3 whether or not there is an abnormal state. In the embodiment, first, the infrared sensor composed of the light projecting portion 11 and the light receiving portion 12 is used to project light. It is determined whether or not the emitted infrared rays are shielded.

On the other hand, the image analysis by the ordinary cameras 18 and 19 is also carried out. In this image analysis, the image data comparison circuit 34 makes the light spots S1 and S2 projected on the regions L1 and L2 shown in FIG. The current image data and the latched previous image data are compared to determine whether or not the light beam spot on the image is shielded. Then, when the light blocking state of the light beam spot is detected and at the same time the light blocking state of the infrared sensor is detected (under the AND condition), the determination circuit 35 determines that the light beam is in an abnormal state.

At this time, the current image data and the latched image data are also similarly compared in the second areas M1 and M2, and the change (increased light) state is determined in the second areas M1 and M2 ( When the light-shielding state is not detected in the light beam spot) and the light-shielding state is determined by the infrared sensor, it is determined to be an abnormal state. Therefore,
In this case, for example, an intruder who wears whitish clothes or clothes whose color is confused with infrared light is easily detected, and cannot be detected by the infrared ray spots S1 and S2. Even an intruder will be reliably detected. Then, at the time of this abnormality, an alarm is output from the speaker 37 by the alarm output in step F4 of FIG.
The images photographed in 8 and 19 are displayed on the monitor of the center or recorded in the recording device.

According to the monitoring utilizing such image analysis, since the cameras 18 and 19 constantly photograph the monitoring area in which the abnormal state is detected, the images at the time of the abnormal state are immediately captured by the cameras 18 and 19. There is an advantage that it is not necessary to move the surveillance camera to the area where the abnormality occurs, unlike the conventional case. Moreover, since the image comparison analysis is performed using the plurality of light beam spots S1 and S2, birds, small animals and the like can be accurately detected without being erroneously detected as an intruder.

Further, in the embodiment, the image of the intruder at the time of abnormality is stored in the frame memory 40, and at night, the strobes 20 and 21 emit light and the intruder illuminated by the flashlight is recorded in the frame memory 40. It Therefore, regardless of whether it is day or night, there is an advantage that the frame memory 40 can efficiently output the image at the moment when the abnormality is detected in real time. That is, even in the presence of security guards and the like, the monitor is not always being observed, and in many cases an intruder at the time of abnormality cannot be confirmed on the monitor. Also, VT
Even when an image is recorded in R or the like, it takes some time to rewind and reproduce the recorded image.

However, in the embodiment, the frame memory 40
Since the one-frame image at the moment of an abnormality is output in real time and the image display can be retained,
It is possible to immediately check the intruder and take prompt action in an emergency situation. It should be noted that the image data recorded in the frame memory 40 is the same as that in step F in FIG.
When the reset operation (manual or automatic) by 5 is performed, the switch circuit 41 switches to the current image.

On the other hand, in step F6 of FIG. 5 described above, it is determined whether or not there is a positional deviation between the light beam spots S1 and S2. If this deviation has occurred, then within the range C1 that can be followed by step F7. It is determined whether or not. That is,
It is judged whether or not the deviation is within the judgment area C1 shown in FIG. 4 (whether it is not a temporary shake). When YES, the spot area L is set at a position such as the light spot SP (FIG. 4). To move. At the same time, the second area M is also moved to the upper side of this spot area L. Therefore, it is possible to maintain a good detection state even when the projection position of the light beam is displaced due to an earthquake, ground deformation, or the like.

In the above embodiment, an example using the inexpensive ordinary cameras 18 and 19 is shown.
A near-infrared camera may be used in place of Nos. 8 and 19, and in this case, photographing at the time of abnormality during nighttime monitoring can be favorably performed. Further, an infrared sensor including the light projecting unit 11 and the light receiving unit 12 may be additionally arranged above the pole-shaped main body 10.

[0037]

As described above, according to the first aspect of the present invention, the light blocking state of a plurality of light beam spots projected at a predetermined interval is determined by image comparison, and an intruder is detected based on the light blocking state. Since the existing abnormal state is detected, only objects of a certain size excluding birds and small animals are detected, and it is possible to reliably observe intruders and the like. Moreover, there is an advantage that an intruder can be directly photographed by the surveillance camera in case of an abnormality.

[0038] In addition, the circumference of the projected light rays spot area
Since the change state is also determined for the black second area of the circle , for example, even when the color of clothes of the intruder is similar to the above-mentioned light ray color, or even when wearing white clothes in the daytime, Presence can be detected well, and high detection accuracy can be maintained.

According to the second aspect of the invention, the position variation of the projected spot is automatically tracked,
Since at least the spot area is automatically reset to a shifted position, there is an advantage that a good detection state can be maintained even if the projection position of the light beam shifts due to an earthquake or ground movement.

According to the third aspect of the invention, since one frame image at that time is stored in the frame memory at the time of abnormality, a real-time image at the moment when an intruder is detected is displayed on a monitor or the like, There is an advantage that quick confirmation and response can be made.

[Brief description of drawings]

FIG. 1 is a block diagram showing a main circuit of a monitoring device using image analysis according to an embodiment of the present invention.

FIG. 2 is a diagram showing a configuration arrangement state of each member of the embodiment.

FIG. 3 is a diagram showing an arrangement state of a light projecting unit and a setting state of a detection area in a screen in the pole-shaped main body of the embodiment.

FIG. 4 is a diagram illustrating a moving state of a light beam spot and a determination area for determining a spot position deviation according to the embodiment.

FIG. 5 is a flowchart showing the main operation of the embodiment.

[Explanation of symbols]

10A, 10B ... Main body, 11, 13, 14, 15, 16 ... Projector, 12 ... Light receiving part, 18, 19 ... Camera, 20, 21 ... Strobe, 22 ... Control circuit, 33 ... Memory (field memory), 34 ... Image data comparison circuit, 35 ... Judgment circuit, 40 ... Frame memory, 45 ... Spot position setting circuit, 46 ... Position shift amount determination circuit.

Continuation of the front page (56) References JP-A-4-175998 (JP, A) JP-A 1-271898 (JP, A) JP-A 7-37064 (JP, A) Actual Kaihei 5-43091 (JP , U) (58) Fields investigated (Int.Cl. ⁷ , DB name) G08B 1/00-15/02

Claims (3)

(57) [Claims] 1. A light projecting portion for projecting a plurality of light rays into a surveillance area at a predetermined interval, and a black second region set in a main body around the light projecting portion.
And a plurality of light beam spots projected from the light projecting section and the above
A surveillance camera for photographing the surveillance area including the second area , and an image of the light spot obtained by the surveillance camera and
An image comparison circuit that sequentially compares the images of the black second region, and a light blocking state and a second state based on the output of the image comparison circuit .
A determination circuit for determining the change state of the area, and the light blocking state of the light ray determined by the determination circuit and the second area
A monitoring device using image analysis configured to detect an abnormal condition in the monitoring area from the changed condition and output the video signal from the monitoring camera to the monitor when the abnormality occurs. 2. Determining the amount of positional deviation of the light beam spot
A position deviation amount judgment circuit is provided, and this position deviation amount judgment circuit
Determines that the spot position has shifted by a predetermined amount.
At least the spots targeted for the image comparison above.
The monitoring device utilizing image analysis according to claim 1, wherein the monitoring area is automatically reset . 3. When the above abnormal condition is detected
A frame memory for storing frame images is provided to
So that the image in the memory can be output to the monitor and held
The monitoring device using the image analysis according to claim 1 or 2 , characterized in that.