JPH05344506A - Method and device for recognizing image abnormality - Google Patents

Abstract

PURPOSE:To selectively extract image abnormality by simple image processing and to detect its cause. CONSTITUTION:A video camera 1 takes a picture for abnormality such as fire in a monitored area. The picture is digitized by an A/D converter 2 and the digital image is processed by an image abnormality detector 3. The detector 3 detects the abnormality of the image and outputs the detected data to a detected data distributor 4. The distributor 4 supplies a control signal S2 to a switch SW. On the other hand, the abnormality in the monitored area is detected by various sensors such as a fire sensor 6 and an illuminance sensor 7. Respective AND circuits A1 to A8 drive their corresponding alarms 14 to 21 for alarming fire detection, smoke generation and illuminance abnormality based upon the control signal S2 outputted from the distributor 4 and the various sensor output signals. Consequently the cause of the abnormal image recognized by the detector 3 can be identified.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an image abnormality recognition method suitable for use in an apparatus for detecting an abnormality in a monitoring area by image processing, and the apparatus.

[0002]

2. Description of the Related Art Conventionally, in public places (banks, shops, etc.) where unspecified people come in and out, and in facilities (power plants, power transmission facilities, etc.) where there is a danger due to some kind of intruder, there are specified places. A video camera is installed in the camera and an image captured from the video camera is monitored to prevent an abnormality from occurring. Further, in such an image abnormality recognizing device, an image abnormality is automatically detected by performing a complicated calculation such as Fourier transform on an image captured from a video camera.

[0003]

By the way, in the above-mentioned conventional image abnormality recognition apparatus, since complicated calculation is required, a system capable of high-speed calculation is indispensable. Therefore, the device causes a problem that the cost is increased. In addition, in the outdoors where the background moves in a complicated manner, it is difficult to distinguish a motion other than a moving object from the above-mentioned image abnormality when a device according to the conventional detection method is used.

Further, even if some abnormality is detected in the image, it is not possible to determine the cause of the abnormality. For example, when a fire occurs, changes in brightness and darkness and changes in interiors such as walls and ceilings are detected as abnormal images in an image captured by a video camera. However, the cause of the abnormality in the image cannot be detected. Therefore, it is not possible to promptly determine how to deal with the result of the image abnormality. As described above, the conventional abnormality recognition device has a problem that it is not possible to detect what causes the change in the image, that is, the abnormality, and what kind of result it causes.

The present invention has been made in view of the above-mentioned problems, and it is possible to selectively extract image abnormalities by simple image processing and also detect the cause thereof. It is an object to provide a recognition method and its device.

[0006]

In order to solve the above-mentioned problems, in the invention according to claim 1, the images for a predetermined period in the past are accumulated with respect to the images sequentially supplied, and the images are stored in the storage means. An image stored in the storage means is averaged, the averaged image is subtracted from the sequentially supplied images, and the subtraction result image and the detection result of at least one detection means for detecting an environmental change It is characterized by issuing an alarm based on and.

According to a second aspect of the present invention, the storage means is provided, and the images sequentially supplied are fetched at predetermined time intervals, and the accumulated images in a predetermined past period are stored in the storage means, and the storage means is provided. Standard image generating means for averaging and outputting the accumulated images output by, and subtracting means for subtracting and outputting the averaged image output by the standard image generating means from the sequentially supplied images, and an environment It is characterized by comprising at least one detecting means for detecting a change, and an alarm generating means for issuing an alarm based on the image output by the subtracting means and the detection result output by the detecting means.

According to a third aspect of the present invention, in the image abnormality recognition method according to the first aspect, the sequentially supplied images are temporarily stored in the current image storage means and also stored in the current image storage means. It is characterized in that the averaged image is subtracted from the image, and an alarm is issued based on the image of the subtraction result and the detection result of at least one detection means for detecting the environmental change.

[0009]

According to the first and second aspects of the invention, the images for a predetermined period in the past are cumulatively stored for the sequentially supplied images and stored in the storage means, and the images stored in the storage means are stored. A standard image means is used for averaging, and the averaged images are subtracted from the sequentially supplied images by a subtracting means, and an alarm is issued based on the image of the subtraction result and the detection result of at least one detecting means for detecting an environmental change. It is characterized by emitting.

According to the invention described in claim 3, in the invention described in claim 1, the images sequentially supplied are stored in the current image storage means, and the standard image is extracted from the images stored in the current image storage means. Is subtracted, and an alarm is issued based on the image of the subtraction result and the detection result of at least one detecting means for detecting the environmental change.

[0011]

Embodiments of the present invention will now be described with reference to the drawings. FIG. 1 is a block diagram showing the configuration of an embodiment of the present invention. In this figure, the video camera 1
For example, if the room is to be monitored, it is installed on the ceiling or wall, and if the room is to be monitored outdoors, it is installed on the outer wall of the building, etc. To the A / D converter 2 and the change-over switch SW at.

The A / D converter 2 converts the video signal S1 output from the video camera 1 into digital video data D1 according to a basic clock signal CL (every time t) described later, and then supplies the digital video data D1 to the image abnormality detecting device 3. To do. The image abnormality detecting device 3 detects the abnormality on the image by taking in the video data D1 for each frame at a predetermined time interval and performing the predetermined image processing, and detects only the abnormal image, the detected data distributor 4 Supply to. The detection data distributor 4 supplies the abnormal image to the monitor M1 and outputs the detection signal S2 notifying that the abnormal image is supplied to the AND circuit A1.
1, A2, ..., A8 on one input end and the switch S
Supply to W. The detailed configuration and operation of the image abnormality detection device 3 will be described later. When the detection signal S2 is supplied, the switch SW supplies the video signal S1 output from the video camera 1 to the monitor M2. Further, the controller 5 performs various controls of the image abnormality detection device 3, and the sampling interval of the video data and the like are set by this.

On the other hand, a fire sensor 6, an illuminance sensor 7, a human sensor 8, a smoke sensor 9, a seismograph 10, a gas sensor 11, a sound pressure sensor 12 and a water infiltration sensor 13 are installed at predetermined locations in the monitoring area. There is. These fire sensor 6, illuminance sensor 7, human sensor 8, smoke sensor 9, seismograph 10, gas sensor 11, sound pressure sensor 12, and infiltration sensor 13
Are supplied to the other input ends of the AND circuits A1 to A8, respectively. In this embodiment, the outputs from the gas sensor 11 and the sound pressure sensor 12 are supplied to the input end of the AND circuit A7. Each AND circuit A1
A8 is adapted to take the logical product of the signals supplied to both input terminals and output the result. That is, when the detection signal S2 is supplied and the outputs from the sensors 6, 7, ..., 13 are supplied to the AND circuits A1 to A8, the outputs are set to “1”. However, AND circuit A
7 outputs "1" only when the outputs of the gas sensor 11 and the sound pressure sensor 12 are supplied. The outputs of the AND circuits A1 to A8 are provided for the alarms 14, 15, 16, 17, 18, 19, 20, 21 provided corresponding to the respective outputs.
Is being supplied to. The notification devices 14 to 21 emit light or sound when the outputs of the corresponding AND circuits A1 to A8 become "1".

Next, the detailed configuration and operation of the above-described image abnormality detecting device 3 will be described with reference to FIGS. 2 to 5. FIG. 2 is a block diagram showing the configuration of the image abnormality detecting device 3 according to this embodiment. Further, FIG. 3 is a conceptual diagram for explaining a method of detecting an image abnormality. First, in FIG. 2, the current frame memory 22 is a storage device that stores the current video data (one frame), and stores the video data output by the A / D converter described above at each time Δt shown in FIG. At the same time, it outputs to the subtractor 25 every time Δt. Further, the standard screen generation circuit 23 is sequentially supplied with the video data output from the A / D converter at each time Δt, as in the current frame memory 22. This standard screen generation circuit 23 generates a standard screen S by averaging n screens for each time Δτ (Δτ ≧ Δt) in the past by time τ with respect to time t. This standard screen S is a background (still object) obtained by removing an abnormal image from the image captured by the video camera.
This is only a screen and serves as a reference for detecting an abnormal image. Here, the generation of the standard screen S and the extraction of the abnormal image will be described.

First, i dot × j at an arbitrary time t
The matrix of the screen composed of dots is represented by Dij (dij, t). Also, the matrix of the desired standard screen S is Sij
(Sij, t). On the other hand, the past n of the time τ described above
The individual screens are Dij (t-τ), Dij (t-τ-Δτ),
Dij (t−τ−2 · Δτ), ………… Dij {t−τ−
It can be represented by (n−1) Δτ}. Then these n
When the average matrix Dij (dij) of the individual data matrices is calculated,

[Equation 1] Becomes This average matrix Dij (dij) has a feature of masking an object that has changed imagewise during the time (n-1) Δτ. That is, if there is no change between each screen (hereinafter referred to as a frame), the average value is used as it is. On the other hand, if there is any change, the image in each frame changes, so the change is averaged. Be masked. Furthermore, even if there is some instantaneous change (brightness change, noise, etc.), it will be averaged,
It is masked by the background.

Therefore, the final mean matrix Dij
(Dij) is at least time {t-τ- (n-1) Δ
From [tau]} to time (t- [tau]), only the background image from which the image changed in the field of view of the video camera is removed.
That is, the present invention is characterized in that the screen including only the background image is the standard screen S. Therefore, the matrix Sij (sij, t) of the standard screen S is

[Equation 2] Becomes Note that the time τ may be “0”, and in that case, the standard screen S is displayed from the time t to the time {t− (n−1).
It is generated based on the past up to Δτ}.

On the standard screen S thus obtained,
It has the following features. (1) The S / N ratio of the standard screen S is improved to the square root times n. Therefore, the random noise in the screen is significantly reduced. (2) The moving objects (or abnormalities) included in each of the n screens in the past from the time {τ + nΔτ} are 1 / n on the standard screen S and thus are almost removed. (3) Rapid changes in screen brightness (flashes of lightning, shading by clouds, etc.) are averaged and removed. Since the standard screen S has the above-mentioned features, it can be used as a reference screen for detecting an abnormal image at an arbitrary time t.

In this way, the standard screen S generated by the standard screen generating circuit 23 is stored in the standard screen memory 24 every time Δτ. The standard screen memory 24 stores the standard screen S and outputs the standard screen S to the subtractor 25 at each time Δτ that is an integral multiple of the time Δt. The subtraction circuit 25 subtracts the standard screen S from the video data output from the current frame memory 22 every time period Δτ. Since the standard screen S is a background screen in the field of view of the video camera as described above, the calculation result of the subtraction circuit 25 is an image in which the background is removed from the image at that time. That is, the subtraction circuit 25 outputs only video data of some change (abnormality) at that time. This video data is stored in the abnormal image memory 26.

Next, the configurations of the standard screen generating circuit 23 and the standard screen memory 24 described above will be described with reference to the block diagram shown in FIG. FIG. 4 shows A according to the present embodiment.
3 is a block diagram showing a configuration of a standard screen generation circuit 23 including A / D converters A1 to A8, a current frame memory 22, and an abnormal image memory 26. FIG. In this figure, parts corresponding to the parts of the block diagram shown in FIG. 1 are assigned the same reference numerals and explanations thereof are omitted. In the figure, the video data output from the A / D converter includes the current frame memory 1 and adders 34-1 to 34-1 to be described later for each basic clock CL signal.
34-n (n; number of frames). Next, the frame memory / address counter 30 generates the address signal AD and the frame signal FL in accordance with the basic clock signal CL, and the current frame memory 22, the frame memory selection counter 31, and the frame memory 35 described later.
-1 to 35-n. Frame memory selection counter 3
1 is stored in the select signals CS1 to CSn for selecting which frame memory 35-1 to 35-n to output the video data in accordance with the frame signal FL and which frame memory 35-1 to 35-n. A control signal CD for selecting whether to output the existing video data is output. The select signals CS1 to CSn are gate circuits 33-1 to 33, respectively.
-n is supplied. Further, the control signal CD is the selector circuit 3
6 is supplied.

Each of the gate circuits 33-1 to 33-n takes the logical product of the video data output from the frame memories 35-1 to 35-n and the select signals CS1 to CSn, and the result is added by the adder 34-1. Output to ~ 34-n. That is, the gate circuit 3
3-1 to 33-n send the video data stored in the frame memories 35-1 to 35-n to the adders 34-1 to 34-n only when the select signals CS1 to CSn become high level. Output. The adders 34-1 to 34-n add the video data and the video data stored in the frame memories 35-1 to 35-n at each timing, and add the frame memory 3
Output to 5-1 to 35-n. Frame memories 35-1 to 35
-n sequentially stores the supplied video data according to the address signal. That is, the frame memories 35-1 to 35-n
In each of the above, the video data of a plurality of frames are overlapped (added) and stored, and the video data of each frame memory is shifted by a time Δτ (details will be described later). ).

Next, the selector circuit 36 controls the control signal CD.
One of the frame memories 35-1 to 35-n according to
The video data stored in one of them is selectively output to the multiplier 37. The multiplier 37 adds a constant K to the selected video data.
Multiply (= 1 / n) and output to the subtractor 38. The subtractor 38 is supplied with the video data at each time t stored in the current frame memory 22, subtracts the video data at each time Δτ from the selected video data, and subtracts the resulting video data. R is the abnormal image memory 26 described above.
Supply to. The abnormal image memory 26 stores the video data R
Is stored and output to the detection data distributor shown in FIG. 1 at intervals of time Δτ.

Next, regarding the operation of this embodiment described above,
This will be described with reference to the timing charts shown in FIGS. 5 (a) to 5 (h). It should be noted that, here, the number of frames used to create the standard screen S is set to 3 for simplification of description. First, the standard screen S is created by inputting a power-on or operation instruction from the controller. First, at time t0, the A / D converter 10 converts the image captured by the video camera into image data A and outputs it to the current frame memory 22 and the adders 34-1 to 34-3. The video data A is stored in the current frame memory 22. Also, the frame memory / address counter 3
0 outputs the above-mentioned address signal AD and frame signal FR according to the basic clock signal CL. Next, the frame memory selection counter 31 causes the frame signal F
According to R, the select signal CS1 is set to low level (see FIG. 5 (b)). Therefore, the gate circuit 53-1 is closed and only the video data A output from the adder 54-1 is stored in the frame memory 55-1 (FIG. 5).
(See (e)). In addition, the frame memories 35-2 and 35
-3 also stores the video data A (see FIGS. 5 (f) and 5 (g)), but when creating the standard screen S for the first time, it is necessary to consider whether or not to add to the previous data. There is no.
Therefore, in this case, the select signals CS2 and CS3
Is at a high level as shown in FIGS. 5 (c) and 5 (d). On the other hand, the current frame memory 22 is sequentially updated from time t0 to time t1 while storing the video data from the video camera every time Δt.

Next, at time t1 after Δτ from time t0, the A / D converter converts the image captured by the video camera into the image data B, and the current frame memory 22 and the adders 34-1 to 34-. Output to n. The video data B is stored in the current frame memory 22 at the time t1. Further, the frame memory / address counter 30
The address signal A described above according to the basic clock signal CL
D and the frame signal FR are output. Next, the frame memory selection counter 31 sets the select signal CS1 to the high level and the select signal CS2 to the low level according to the frame signal FR (see FIGS. 5B and 5C). Therefore, the gate circuit 53-1 is opened, and the video data A stored in the frame memory 55-1 and the video data B are added in the adder 54-1 (referred to as video data (A + B)). .. This video data (A + B) is stored in the frame memory 35-1 (see FIG. 5 (e)). Since the gate circuit 33-2 is closed, the frame memory 35-2 has an adder 34
The video data B is stored as it is via -2 (Fig. 5
(See (f)).

Next, at time t2 after Δτ from time t1, the A / D converter converts the video captured by the camera into video data C and outputs it to the current frame memory 22 and the adder 34. The video data C at time t2 is stored in the current frame memory 22. Next, the frame memory selection counter 31 sets the select signal CS2 to the high level and the select signal CS3 to the low level according to the frame signal FL (FIGS. 5C and 5).
(See (d)). Since the select signal CS1 maintains the high level state, the gate circuit 33-1 is opened, and the adder 34-1 adds the video data (A + B) and the video data C (video data ( A + B + C)
And). Also, the gate circuit 33-2 is opened,
In the adder 34-2, the video data B stored in the frame memory 35-2 and the video data C are added (video data (B + C)). This video data (B
+ C) is stored in the frame memory 35-2 (FIG. 4).
(See (f)). Further, since the gate circuit 33-3 is closed, the video data C is stored as it is in the frame memory 35-3 via the adder 34-3 (see FIG. 5 (g)).

As described above, the video data A, B and C output by the video camera 1 for each Δτ are first recorded in the frame memory 3
It is stored in 5-1. Then, at time t3 when Δτ has elapsed, the video data D is stored in the current frame memory 22 (see FIG. 5 (e)). The frame memory selection counter 31 selects the output data of the frame memory 35-1 by the control signal CD. As a result, the video data (A + B +) stored in the frame memory 35-1
C) is supplied to the multiplier 37 via the selector circuit 36. Then, in the multiplier 37, the multiplier K (= 1 / n)
Is subtracted as a standard screen S averaged by multiplying
8 are supplied. In the subtractor 38, the current frame memory 1
From the video data D stored in the above video data (A
+ B + C) is subtracted (see FIG. 5 (h)). in this case,
If the video data D includes an abnormal image, the subtracter 38
The video data R output by is the video data of only the abnormal image. The video data R is stored in the abnormal image memory 26.

On the other hand, at the same time t3, the frame memory selection counter 31 follows the frame signal FR.
The select signal CS1 is set to low level, and the select signal C
Set S3 to high level (Fig. 5 (b) and Fig. 5 (d))
reference). The select signal CS2 maintains the high level state (see FIG. 5 (c)). Therefore, the gate circuit 33-1 is closed and only the video data D is stored in the frame memory 35-1 as shown in FIG.
Further, since the gate circuit 33-2 is opened, the frame memory 35-2 is newly updated with image data (B + C + D) obtained by adding the image data (B + C) stored in the frame memory and the image data D. Stored in (Fig. 5
(See (f)). Furthermore, since the gate circuit 33-3 is open, the video data (C
+ D) is stored (see FIG. 5 (g)).

Next, at time t4, the frame memory selection counter 31 selects the output data of the frame memory 35-2 by the control signal CD. As a result, the video data (B + C +) stored in the frame memory 35-2
D) is supplied to the multiplier 37 via the selector circuit 36. Then, in the multiplier 37, the video data (B + C
Multiplier K (= 1 / n) is multiplied by + D), and the subtracter 3 is used as an averaged standard screen S (= (B + C + D) ').
8 are supplied. In the subtractor 38, the current frame memory 2
The video data (B + C + D) 'is subtracted from the video data E stored in No. 2 (see FIG. 5 (h)). In this case, if the video data E includes an abnormal image, the video data R output by the subtractor 38 is video data of only the abnormal image. This video data R is stored in the abnormal image memory 26.
Memorized in. On the other hand, at the same time t4, the image data (D + E), the image data E, and the image data (C + D + E) are stored in the frame memories 35-1, 35-2 and 35-3, respectively.
Is stored (see FIG. 5 (e) and FIG. 5 (g)).

Then, at time t5, the video data (C + D + E) stored in the frame memory 35-3.
Is multiplied by a multiplier K (= 1 / n) to obtain an averaged standard screen S (= (C + D + E) ′). And FIG.
As shown in (h), from the video data F stored in the current frame memory 22, the video data (C + D +
E) ′ is subtracted and stored in the abnormal image memory 26 as the video data R. Then, the frame memory 35-1,
The video data (D + E +
F), video data (E + F), and video data F are stored (see FIGS. 5 (e) and 5 (f)).

Thereafter, even after time t6, the video data (standard screen S) stored in the frame memories 35-1, 35-2 and 35-3 are sequentially supplied to the subtractor 38, and new video data R is generated. It is obtained and stored in the abnormal image memory 26. Also, the video data G, H, ... at each time.
Are stored in the frame memories 35-1, 35-2 and 35-3 while being sequentially shifted by 3 frames and shifted by 1 frame. Then, at each time after the time t3, the video data R stored in the abnormal image memory 26 is sequentially supplied to the detection data distributor shown in FIG. As described above, in the present embodiment, the standard screen S obtained by averaging a plurality of frames in the past is created, and when a moving object enters the field of view of the camera, based on the standard screen S, only that moving object is displayed. The image is taken out and stored in the abnormal image memory 26.

Next, the operation of the above-described abnormal image recognition apparatus will be described with reference to some examples when various abnormalities occur in the monitoring area. For example, when a fire occurs in the monitoring area, the video camera 1 stores the flame due to the fire and the change in the brightness of the image due to the flame as an abnormal image in the abnormal image memory 26, and the detected data distributor 4 Is output to. The detection data distributor 4 recognizes that some abnormality has occurred due to the supply of the abnormal image, supplies the control signal S2 to the switch SW, and the video data stored in the abnormal image memory 26 on the monitor M1. To supply. As a result, the image of the fire scene captured by the video camera 1 is displayed on the monitor M2, and the other monitor M1 is displayed.
The abnormal image stored in the moving memory 26 is displayed on the screen. On the other hand, the output signals from the fire sensor 6, the illuminance sensor 7, and the smoke sensor 9 are supplied to the other input ends of the AND circuits A1, A2, A4 by heat and smoke from the fire. Therefore, the outputs of the AND circuits A1, A2, A4 become "1", and the alarms 14, 15, 17 for notifying the fire detection, smoke generation, and illuminance abnormality operate. As a result, it is identified that the abnormal image recognized by the abnormal image detection device 3 is caused by a fire.

When a power failure occurs in the monitoring area, the video camera 1 stores the change in brightness of the image due to the power failure in the abnormal image memory 26 as an abnormal image and outputs it to the detection data distributor 4. It The detection data distributor 4 is supplied with the abnormal image,
Upon recognizing that some abnormality has occurred, the control signal S2 is supplied to the switch SW and the video data stored in the abnormal image memory 26 is supplied to the monitor M1. As a result, the image of the surveillance area photographed by the video camera 1 is displayed on the monitor M2, and the other monitor M is displayed.
An abnormal image stored in the moving memory 26 is displayed at 1. On the other hand, when a power failure occurs, the output signal from the illuminance sensor 7 is supplied to the other input terminal of the AND circuit A2. Therefore, the output of the AND circuit A2 becomes "1", and the alarm device 15 for notifying the illuminance abnormality operates. As a result, it is identified that the abnormal image recognized by the abnormal image detecting device 3 is caused by the illuminance abnormality (in this case, power failure).

Further, when a gas leak occurs in the monitoring area, and it ignites and explodes, a change in the image due to the explosion (change in brightness and darkness, change in shape of the building) is abnormal in the video camera 1. The image is stored in the abnormal image memory 26 as an image and is output to the detection data distributor 4. The detection data distributor 4 recognizes that some abnormality has occurred due to the supply of the abnormal image, and the control signal S
2 is supplied to the switch SW and the video data stored in the abnormal image memory 26 is supplied to the monitor M1.
As a result, the monitor M2 displays the image of the monitoring area captured by the video camera, and the other monitor M1 displays the abnormal image stored in the moving memory 26. On the other hand, when an explosion occurs, the output signals from the gas sensor 11 and the sound pressure sensor 12 are supplied to the other input ends of the AND circuits A6 and A7. Therefore, the outputs of the AND circuits A6 and A7 become "1", and the alarms 19 and 20 for notifying the gas explosion and the abnormal sound are activated. As a result, it can be identified that the abnormal image recognized by the abnormal image detection device 3 is caused by a gas explosion (in this case, a power failure).

In addition to what has been described above, even when a person enters the surveillance area, the person detection / informing device 16 operates by the output signal from the person sensor 8, so that the abnormal image photographed by the video camera 1 is not detected. It can be identified as the result of the intrusion. Further, even when an image is taken by the video camera 1 and an abnormal image is taken due to an earthquake, the output signal from the seismograph 10 is supplied to the AND circuit A5, so that the above-mentioned abnormality occurs. It can be identified that the image was caused by an earthquake.
Further, even when the image taken by the video camera 1 is greatly changed due to the inundation and an abnormal image is taken, the output signal from the inundation sensor 13 is the AND circuit A.
It is possible to identify that the abnormal image is caused by the inundation because it is supplied to No. 8.

In the above embodiment, the frame address counter 30, the frame memory selection counter 31,
Although the gate circuits 33-1 to 33-n, the selector circuit 36, the multiplier 37, the subtractor 38, and the like are configured by hardware, they may be implemented by software. The moving object described above is not limited to the type or state of the substance forming the moving object, and may be any object having a light-dark contrast with respect to the background. Further, the above-described video camera 1 is not limited to being sensitive to only visible light,
For example, it may be a means that is sensitive to invisible light rays (ultraviolet rays, infrared rays, etc.), X-rays or laser rays.

[0035]

As described above, according to the present invention, the above-mentioned images for a predetermined period in the past are cumulatively stored in the storage means and sequentially stored in the storage means, and stored in the storage means. The averaged image is averaged by the standard image means, the averaged image is subtracted from the sequentially supplied image by the subtracting means, and the subtraction result image and the detection result of at least one detecting means for detecting the environmental change are obtained. Since the alarm is issued based on the above, there is an advantage that the abnormality of the image can be selectively extracted by the simple image processing and the cause thereof can be detected.

[Brief description of drawings]

FIG. 1 is a block diagram showing a configuration of an exemplary embodiment of the present invention.

FIG. 2 is a block diagram showing a configuration of an image abnormality detection apparatus of the same embodiment.

FIG. 3 is a time chart for explaining the operation of the embodiment.

FIG. 4 is a block diagram showing a detailed configuration of an image abnormality detecting apparatus according to the present exemplary embodiment.

5A to 5H are timing charts for explaining the operation of the present embodiment, FIG. 5A being video data,
(B) is the select signal CS1, (c) is the select signal C
S2, (d) is the select signal CS3, (e) is the output of the frame memory 35-1, (f) is the output of the frame memory 35-2, (g) is the output of the frame memory 35-3, and (h) is 6 is a diagram showing an output of the subtractor 38. FIG.

[Explanation of symbols]

 6 Fire Sensor (Detecting Means) 7 Illuminance Sensor (Detecting Means) 8 Human Sensor (Detecting Means) 9 Smoke Sensor (Detecting Means) 10 Seismometer (Detecting Means) 11 Gas Sensor (Detecting Means) 12 Sound Pressure Sensor (Detecting Means) 13 Water immersion sensor (detection means) 14 to 21 Alarm device (alarm generation means) 22 Current frame memory (current image storage means) 23 Standard screen generation circuit (standard image generation means) 24 Standard screen memory (storage means) 25 Subtractor (subtraction Means) 26 abnormal image memory

 ─────────────────────────────────────────────────── ─── Continuation of the front page (72) Inventor Katsuhiko Ito 1-5-1 Kiba, Koto-ku, Tokyo Within Fujikura Electric Wire Co., Ltd. (72) Inventor Toru Wada 1-5-1 Kiba, Koto-ku, Tokyo Fujikura Electric Wire Co., Ltd.

Claims (3)

[Claims] 1. A method for accumulating the images in a predetermined period in the past for sequentially supplied images and storing the accumulated images in a storage means,
The images stored in the storage unit are averaged, the averaged image is subtracted from the sequentially supplied images, and the image of the subtraction result and the detection result of at least one detection unit that detects an environmental change are used. A method for recognizing an image abnormality, which is characterized by issuing an alarm. 2. A storage means is provided, images sequentially supplied are taken in at a predetermined time interval, images accumulated in a predetermined period in the past are stored in the storage means, and output by the storage means. Standard image generation means for averaging and outputting images, subtraction means for subtracting and outputting the averaged images output by the standard image generation means from the sequentially supplied images, and at least 1 for detecting an environmental change An image abnormality recognition device comprising: one detection means; and an alarm generation means for issuing an alarm based on the image output by the subtraction means and the detection result output by the detection means. 3. The sequentially supplied images are temporarily stored in a current image storage means, the averaged image is subtracted from the image stored in the current image storage means, and the subtraction result is obtained. The image abnormality recognition method according to claim 1, wherein an alarm is issued based on the image and a detection result of at least one detection unit that detects an environmental change.