JPH0955932A - Abnormality monitoring device abnormality detection method - Google Patents

Abstract

(57) [Abstract] [PROBLEMS] To detect an abnormality of a subject having an unspecified shape. SOLUTION: A normal still image of the monitoring target 1 photographed by the television camera 4 and a photographed still image photographed at the time of monitoring are compared by the CPU 8, and when there is no match, it is considered that an abnormality has occurred, and the multimedia alarm device 9 is provided. Generate an alarm from.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an abnormality detecting method for an abnormality monitoring apparatus which photographs a subject to be monitored and judges the presence or absence of abnormality in the photographed result.

[0002]

2. Description of the Related Art There is an urgent need to monitor the presence or absence of abnormalities in objects in various fields such as plants, equipment, factories, shops, finance, railways, and roads. The types of anomalies include people, cars,
There is an approach of ships, animals, etc., abnormal shape, damage, heat generation, abnormal temperature, etc.

An nuclide has been proposed as an anomaly monitoring device in which an object to be monitored is photographed by a video camera, the photographed result is compared with a prepared normal pattern, and the presence or absence of an abnormality is automatically determined based on the comparison result.

[0004]

Conventionally, in this kind of abnormality monitoring apparatus, a contour image is often extracted from a photographed image as a normal pattern. For example, continuous pixels that have a large difference in luminance value between adjacent pixels with the contour image, that is, a line segment is detected, and the position information of the closed partition area surrounded by this line segment is set as a normal pattern. . However,
In this way, the method of detecting the contour line is time-consuming and
It has a drawback that the process itself is complicated. Therefore, monitoring of suspicious persons such as those whose shapes cannot be specified in advance,
For example, it is not suitable for monitoring for the purpose of theft prevention, and in the past, humans had no choice but to visually check a photographed image on a monitor or the like.

Therefore, it is an object of the present invention to provide an abnormality detecting method for an abnormality monitoring device which is suitable for detecting an abnormal object whose shape cannot be specified and which can be used for general purposes.

[0006]

In order to achieve such an object, the invention of claim 1 provides a normal still image in a normal state of a monitoring target by a camera which converts an image of a subject into an electric signal and outputs the electric signal. Acquired, the monitoring target is captured by the camera to acquire the captured still image at regular intervals,
The monitoring device is characterized in that the normal still image is stored, the photographed still image is compared with the normal still image, and when an unmatched comparison result is obtained, an abnormality warning is given.

According to a second aspect of the present invention, in addition to the first aspect, the monitoring device uses the level value of the electric signal output from the camera for comparison between the photographed still image and the normal still image. Is characterized by.

According to a third aspect of the present invention, in addition to the first aspect of the invention, either one of the photographed still image and the normal still image is displayed with respect to a time-series change of the monitoring target in a normal state. It is characterized by correction.

According to a fourth aspect of the invention, in addition to the first aspect of the invention, a monitor system for receiving the warning of the abnormality and the captured still image is connected to the monitoring device, and the monitoring device acquires at a constant cycle. The captured still image is compressed and transmitted to the monitor system, and the monitor system records the compressed captured still image.

In addition to the invention of claim 1, the invention of claim 5 is characterized in that an infrared camera is used as the camera.

[0011]

BEST MODE FOR CARRYING OUT THE INVENTION Embodiments of the present invention will be described in detail below with reference to the drawings.

FIG. 1 shows a monitoring system to which the present invention is applied. In FIG. 1, the monitoring device 100 and the monitor system 2
00 is connected by a communication network. The television camera 4 images the monitoring target 1 and outputs an analog image signal obtained as a result of the imaging. The analog image signal is amplified by the amplifier 5 and then converted into a digital signal by the analog / digital (A / D) converter 6.
This digital image signal is sent to the CPU via the memory 7.
It is handed over to 8 and used to determine whether there is an abnormality.

More specifically, a still image in which the monitoring target 1 is in a normal state is captured, image features described below are extracted, and a normal pattern is created. Further, at the time of monitoring, a monitoring image signal of the monitoring target 1 is fetched and the image is compared with a normal pattern. The presence or absence of abnormality of the monitoring target 1 is determined based on the comparison result. The determination result and the monitor image signal are sent to the multimedia alarm device 9, displayed on the display device 10, and the abnormality is output from the voice device 11 as a voice message. In addition, the monitor system 200 is provided via the communication device 12.
The monitor image signal and the result of the presence / absence determination of abnormality are sent to the display device 16, and the monitor image is displayed on the display device 16. The voice device 17 outputs a voice message indicating that there is an abnormality. Further, the time information at this time, the monitoring image, and the like are recorded in the recording device 19.

A method of determining the presence / absence of an abnormality in this embodiment will be described. An imaging screen of the monitoring target object 1 in the absence of an abnormal object, for example, a thief is captured M times in a cycle T1. The CPU 8 calculates the average of the level values of the image signal for the same pixel, and the image of one screen constituted by the average value is set as the normal pattern (normal still image of the present invention) in the memory 7
Save it inside.

At the time of monitoring, a monitoring image is captured at a cycle T2. The CPU 8 calculates the total difference between the level value of the monitor image and the level value of the normal pattern for the same pixel. In this calculation, the same difference value can be obtained by summing the level values of all the pixels of the normal pattern and calculating the difference between this sum and the sum of the level values of all the pixels of the monitoring image. When an abnormal object is mixed in the monitoring image, the image signal distribution becomes different from the normal pattern. As a result, since the above-mentioned difference value exceeds the allowable range, both images do not match and an abnormality can be detected. The processing procedure of the CPU 8 on the monitoring apparatus 100 side is shown in FIG. 2 for reference.

In addition to this embodiment, the following example can be implemented.

1) When it is desired to further improve the accuracy of detecting an abnormality, the following processing may be performed.

(A) A difference value for each pixel between the normal pattern and the monitor image is calculated, and the CPU 8 counts the number of pixels for which the difference value exceeds the allowable range. When the count result exceeds a certain number, it is determined that an abnormality has occurred.

(B) When monitoring for 24 hours, a change in brightness appears due to changes in illumination and environment even in a monitor image in a normal state. Therefore, the CPU 8 measures the time, and corrects the image signal level value (may be the total value) of either the normal pattern or the monitor image in accordance with the time. Specifically, a table of correction coefficients corresponding to time is prepared, an appropriate coefficient is acquired by referring to this table, and the coefficient is multiplied by the level value of the image signal. By performing such processing, sunshine change,
It can respond to changes in lighting and power supply fluctuations. Also, 1
It is also possible to take a photograph of the monitoring target every time and use the photographed still image as a normal pattern. In this case, the normal pattern is properly used according to the monitoring time.

(C) In a photographed image, a place of particular interest, for example, for the purpose of theft prevention, an image of only a specific region such as a front door, a safe, etc. is set as an object of determination of abnormality. You can In this case, a pointing device may be connected to the CPU 8 and a specific area of the image displayed on the display device 10 may be variably set by the pointing device.

Further, the detection accuracy of the abnormality may be made different between this specific place and other places.

2) When the monitor image is recorded in the monitor system 200 for a long time, the monitor device 100 may compress the monitor image data and send it to the monitor system 200.

3) As the photographing means, a TV camera for photoelectric conversion is used in this embodiment, but an infrared TV camera may be used to detect an abnormality by the temperature change of the monitored object.

4) The abnormality type message can be output by voice corresponding to the abnormality occurrence pattern. For example, when the facility is destroyed or damaged, after the abnormality of the captured image is detected at a certain point, the location of the abnormality (the pixel position where the brightness change occurs) does not change in the subsequent captured images. . In a moving object such as a thief, the location of the abnormal place changes. Therefore, when an abnormality is detected for the first time, the detected pixel position is stored in the memory 7 by the CPU 8 and compared with the next abnormality detection position. If a match is found, the monitored object is damaged, and if a mismatch is found, the suspicious person and the abnormal content can be identified.

[0025]

As described above, according to the invention of claim 1, the abnormality of the object can be automatically detected by the simple process of comparing the still images, the purpose of preventing theft and the purpose of monitoring the facility. The present invention can be carried out.

According to the second aspect of the present invention, since the level value of the output signal of the camera is used for comparison of still images, the comparison process is simpler and the processing time is shorter.

According to the third aspect of the invention, since the correction is made for the change of the photographing environment, the abnormality detection accuracy is improved.

According to the invention of claim 4, the image obtained as a result of the monitoring is compressed and transferred, so that the monitor system can display the image in real time and can record for a long time.

The invention of claim 5 is suitable for nighttime monitoring, and in particular, lighting equipment is unnecessary.

[Brief description of the drawings]

FIG. 1 is a block diagram showing a configuration of a monitoring system to which the present invention is applied.

FIG. 2 is a flowchart showing an operation procedure of the embodiment of the present invention.

[Explanation of symbols]

 1 Monitoring Target 2 Abnormal Person 3 Illuminator 4 Television Camera 5 Amplifier 6 A / D 7 Memory 8 CPU 9 Multimedia Alarm Device 10, 16 Display Device 11, 17 Audio Device 12, 14 Communication Device 13 Network 100 Monitoring Device 200 Monitoring System

─────────────────────────────────────────────────── ─── Continuation of the front page (51) Int.Cl. ⁶ Identification code Internal reference number FI technical display location G08B 25/00 510 9061-5H G06F 15/70 455B

Claims (5)

[Claims] 1. A normal still image in which a monitoring target is in a normal state is acquired by a camera which converts an image of a subject into an electric signal and is output, the monitoring target is photographed by the camera, and the captured still image is taken at regular intervals. The monitoring device stores the normal still image, compares the captured still image with the normal still image, and warns of an abnormality when a mismatched comparison result is obtained. Anomaly detection method for anomaly monitoring device. 2. The monitoring device uses the level value of an electric signal output from the camera for comparison between the photographed still image and the normal still image.
An abnormality detection method for an abnormality monitoring device as described in. 3. The one of the photographed still image and the normal still image is corrected for a time-series change in a normal state of the monitoring target.
An abnormality detection method for an abnormality monitoring device as described in. 4. The monitor device is connected to a monitor system for receiving the warning of the abnormality and the photographed still image, and the monitor device compresses the photographed still image acquired in a constant cycle and transmits the compressed still image to the monitor system. The abnormality detection method of the abnormality monitoring device according to claim 1, wherein the monitor system records the compressed still image. 5. The abnormality detecting method for an abnormality monitoring device according to claim 1, wherein an infrared camera is used as the camera.