JPH01297793A - Abnormality monitoring device - Google Patents

Abstract

PURPOSE:To reduce erroneous information by extracting the edge of an image representing a monitoring area, counting the number of edges facing in the same direction at every direction, storing the number of edges in the same direction of a reference image in advance, comparing the number of edges in the same direction of the present image with that of edges in the same direction of the reference image, and deciding abnormality. CONSTITUTION:The image at a normal state without having the abnormality such as the invasion of an invader, etc., in the monitoring area is photographed in advance, and the edge is extracted by an edge extraction means 3 by applying A/D conversion. The number of edges in each of directions (1)-(8) is counted by a same direction edge number counting means 4. A count result is stored in the directional table 5 of the reference image. Similarly in case of the reference image, the number of edges in the same direction at every direction for the present image is found, and it is stored in the directional table 6 of the present image. The number of edges in the same direction of the reference image is compared with that of edges in the same direction of the present image by a differential means 7, and the degree of coincidence of the number of edges at each of the directions (1)-(8) is found. Then, the presence/absence of the abnormality is decided by an abnormality decision means 8.

Description

DETAILED DESCRIPTION OF THE INVENTION [Field of Industrial Application] The present invention relates to an abnormality monitoring device that uses an imaging means such as a television camera to detect whether or not an abnormality has occurred in an area to be monitored.

[Conventional technology]

Conventionally, this type of apparatus stores a current image obtained by an imaging means 1 and an image input means 2 and a reference image stored in advance in a current image memory 21 and a reference image memory 20, as shown in FIG.
The difference means 22 calculates the luminance difference between each pixel between the current image and the reference image, and the binarization means 23 binarizes it with a predetermined darkness value. There is a method in which the determination means 8 determines that an abnormality occurs when the number of changed pixels exceeds a set number, and the abnormality output means 9 outputs the result (Japanese Patent Laid-Open No. 60-7593).

[Problem to be solved by the invention]

However, in such a device, when a change in brightness occurs in the monitoring area due to a phenomenon that is not inherently abnormal, such as a change in illuminance due to sunlight, the brightness of the current image changes, so the difference means 22 calculates a constant value. The problem is that when the difference from the reference image is taken, the abnormality determining means 8 determines that the number of changed pixels is larger than the set number and is abnormal, resulting in a false alarm! there were.

The present invention has been made in view of the above points, and an object of the present invention is to provide an abnormality monitoring device that reduces false alarms.

[Means to solve the problem]

The above problem is that if you simply obtain the difference image, the difference between the reference image and the current image will increase due to phenomena that are not abnormal in nature, such as changes in lighting or changes in the amount of light coming in from a window, which will cause an alarm to be issued. It's put away. This occurs simply because the brightness of the images is compared. That is, this occurs because it is not possible to distinguish between changes in image brightness caused by an abnormality such as an intruder's invasion or a fire outbreak, and changes in image brightness caused by changes in sunlight or lighting.

The present invention focuses on the fact that the shape of an object does not change even if the illuminance changes, and uses edge information of an image that is considered to best express the shape of the object. and an image input means for A/D converting the image signal output from the imaging means; and an edge extraction means for extracting the edges of the image from the A/D converted image signal; a reference image direction table that stores the number of edges in the same direction with respect to a reference image indicating a normal state of the monitoring area; An abnormality that determines whether or not there is an abnormality by comparing a current image direction table that stores the number of edges in the same direction for an image, the number of edges in the same direction for the reference image, and the number of edges in the same direction for the current image for each direction. The present invention is characterized in that it comprises a determination means.

[Effect]

An edge extraction means extracts edges of an image representing a monitoring area, and a same direction edge count counting means counts the number of edges facing in the same direction for each direction. The number of edges in the same direction of the reference image is memorized in advance,
The number of edges in the same direction of the current image is compared with the number of edges in the same direction of the reference image, and the presence or absence of an abnormality is determined by the abnormality determining means based on the comparison result.

〔Example〕

DESCRIPTION OF THE PREFERRED EMBODIMENTS The present invention will be explained below based on drawings showing one embodiment of the present invention.

FIG. 1 is a block diagram showing one embodiment of the present invention.

Reference numeral 1 denotes an imaging means such as a television camera, which takes an image of the monitoring area and outputs an image signal. Reference numeral 2 denotes an image input means, which obtains a digital image signal by A/D converting the image signal. Reference numeral 3 denotes edge extraction means, which extracts the edges of the image captured by the image capturing means 1, and can be realized by, for example, a 3×3 filter such as a Sot+el operator.

The S, obel operator is a means often used when extracting edges. If the brightness at the coordinates (x, y) is I (x, y), the brightness of the edge image J (
x, y) can be calculated using the following formula.

J Cx, y) = l I Cx-1, y-1) +
I (x-1.7)+I (X-1,X+1)-
I (X+1.y-1) -I (x11,y)-1
(x+l, X+1)l 10l I (x-1,y-1)+I (x, y-1
)+■ (X+1. y-1)-I (x-L X+
1) I (x.

X+1)-1(χ+1.X+1)1 Furthermore, the extracted edge image is binarized using a predetermined darkness value.

4 is a means for counting the number of edges in the same direction, and an edge extracting means 3
From the edge information obtained, the number of edges in each of the eight directions, for example, ■ to ■, as shown in FIG. 2, is counted.

The type of this direction is set in advance. Here, for example,
The distinction between directions ■ and ■ is determined by which side of the edge is brighter. Furthermore, the number of edges that appear as one on the screen is determined to be one edge for a certain number of pixels (for example, one pixel).

Reference numeral 5 denotes a reference image direction table which stores the number of edges in the same direction with respect to the reference image indicating a normal state of the monitoring area.

Reference numeral 6 denotes a direction table for the current image, which stores the number of upward turns in the same direction each time for the current image that is sequentially input.

7 is a difference means which compares the number of edges in the same direction output from the direction table 5 of the reference image and the number of edges in the same direction output from the direction table 6 of the current image for each corner direction.

Reference numeral 8 denotes an abnormality determining means, which outputs a signal indicating that there is an abnormality when it is determined that there is a significant difference based on the comparison result of the difference means 7. , 9 is an abnormality alarm means, which issues an alarm in response to a signal from the abnormality determination means 8 indicating that there is an abnormality.

Next, the operation of this embodiment will be explained.

First, an image in a normal state in which there is no abnormality such as an intrusion of an intruder into the monitoring area is taken in advance by the imaging means, and the image is converted into A/D by the image input means 2. Edges are extracted by edge extraction means 3 from the A/D converted image. From the extracted edge information, the number of edges in the same direction is counted by means 4 for counting the number of edges in the same direction in each direction from the direction (1) to the force direction. This count result is stored in the reference image direction table 5 as the number of edges in the same direction of the reference image.

After this preparation, the actual monitoring state is entered. Image signals of the current image are sequentially inputted from the imaging means 1, and in the same manner as in the case of the reference image, the number of edges in the same direction in each direction with respect to the current image is determined and stored in the direction table 6 of the current image.

The number of edges in the same direction of the reference image and the number of edges in the same direction of the current image are compared by the difference means 7, and the degree of coincidence of the number of edges is determined for each of the angular directions (1) to (2).

The presence or absence of an abnormality is determined by the abnormality determining means 8 based on the degree of coincidence of the number of edges in each direction (1) to (4) determined by the difference means 7. For example, as shown in FIG. 4, when there is an intruder M, the number of edges in the same direction in each direction with respect to the reference image and the number of edges in the same direction in each direction with respect to the current image are naturally bounded. It is determined that there is an abnormality when the degree of difference exceeds a predetermined threshold (i (value indicating how many lines differ in how many directions).

In response to the output from the abnormality determining means 8 indicating that there is an abnormality, the abnormality warning means issues an alarm to the outside.

FIG. 5 is a block diagram showing another embodiment according to the present invention. In the embodiment shown in FIG.
In front of the reference image memory 10, current image memory 11, difference means 12, difference area extraction means 14, primary abnormality determination means 1
5 and a changeover switch 13 are inserted.

First, a reference image showing a normal state is stored in the reference image memory 10. Next, the changeover switch 13 is switched to the edge extraction means 3 side, and the number of edges in the same direction for each direction with respect to the reference image is stored in the direction table 5 of the reference image using the method described above.

However, in this case, as shown in FIG. 6, the number of edges in the same direction in each direction is stored for each partial region R obtained by dividing the screen.

Next, the changeover switch 13 is switched to the difference means 12 side. The current image taken by the imaging means 1 in the monitoring state is
The difference in brightness between the current image memory 11 and the reference image stored in the reference image memory 10 is calculated, and the difference area extracting means 14 extracts an area with a significant difference. That is, the difference images are averaged for each partial region R, and regions exceeding a preset darkness value are extracted as difference regions.

For example, if there is an intruder M as shown in Figure 3,
The change area S indicated by diagonal lines in the figure is extracted as a difference area.

Next, based on the presence or absence of the difference area, the primary determination means 15 determines whether there is a possibility that an abnormality exists. In other words, through high-speed processing that simply looks at the difference in brightness,
A primary judgment is made.

If it is determined that there is a possibility of abnormality in the first determination,
Switch the changeover switch 13 to the edge extraction means 3 side.

In this case, the output from the primary determining means 15 may be used to automatically switch the changeover switch 13.

Here, a region that has not been extracted as a difference region is given as a mask from the difference region extraction means 14 to the edge extraction means 3, and edges only in the changed region S shown in FIG. 7 are extracted. This is because processing by the edge extraction means 3 generally takes time, so by masking unnecessary parts, the processing speed can be increased.

From the output of the edge extraction means 3, the number of same direction edges is counted by the same direction edge number counting means 4 and is stored in the direction table 6 of the current image. The difference means 7 compares the number of edges in the same direction in the reference image and the number of edges in the same direction in the current image for each change region S, and the abnormality determination means 8 determines whether or not there is an abnormality.

Only when the abnormality determining means 8 also determines that there is an abnormality, the abnormality warning means 9 issues an alarm.

〔Effect of the invention〕

As described above, according to the abnormality monitoring device of the present invention, there is provided an imaging means for imaging a monitoring area, an image input means for A/D converting an image signal output from the imaging means, and an A/D converted image. an edge extracting means for extracting the edges of the image from the signal; a same direction edge count means for counting the number of edges facing in the same direction as the edge extracted by the edge extracting means; A reference image direction table that stores the number of edges in the same direction with respect to a reference image that indicates a state, a current image direction table that stores the number of edges in the same direction with respect to the current image, and a number of edges in the same direction with respect to the reference image and the number of edges in the same direction with respect to the current image. Since the abnormality determining means for determining whether or not there is an abnormality by comparing the numbers in each direction, it is possible to provide an abnormality monitoring device that reduces false alarms.

[Brief explanation of the drawing]

FIG. 1 is a block diagram showing an embodiment of the present invention, FIG. 2 is a schematic diagram showing the edge direction for explaining the operation of the above, FIG. 3 is a schematic diagram showing the screen in a normal state, and FIG. 4 is a schematic diagram showing a screen in an abnormal state, FIG. 5 is a block diagram showing another embodiment of the present invention, FIG. 6 is a schematic diagram of a screen for explaining the operation of the above, and FIG. 7 is a schematic diagram showing a screen in an abnormal state. 8 is a schematic diagram of a screen for explaining the operation of the same as above, and FIG. 8 is a block diagram showing a conventional example. 1-Imaging means 2-Image input means 3-Edge extraction means 4-Same direction edge number counting means 5-Reference image direction table 6-Current image direction table 7--Difference means 8--Anomaly determination means 9- Abnormal alarm means patent applicant Matsushita Electric Works Co., Ltd.

Claims (1)

[Claims] (1) an imaging means for imaging a monitoring area; an image input means for A/D converting an image signal output from the imaging means;
an edge extracting means for extracting image edges from an A/D converted image signal; and a same direction edge number counting means for counting the number of edges facing in the same direction as the edge extracted by the edge extracting means. , a reference image direction table that stores the number of edges in the same direction for a reference image that indicates a normal state of the monitoring area, a current image direction table that stores the number of edges in the same direction for the current image, and a table that stores the number of edges in the same direction for the reference image and the current An abnormality monitoring device comprising abnormality determining means for determining whether or not there is an abnormality by comparing the number of edges in the same direction with respect to an image in each direction.