JP2503613B2 - Abnormality monitoring device - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION [Industrial field of use] The present invention uses an image pickup means such as a television camera to digitally process an image signal of the image pickup means and detect whether there is an abnormality such as an intruder in the surveillance area. The present invention relates to a monitoring device.

[Conventional technology]

Conventionally, as this type of device, there is one called a motion detector. This is to detect the brightness level due to the movement of a person, an object, or the like in the monitoring area by taking the difference between two consecutive images for about 16 images per second (change detection method). However, in this case, if a moving object or the like is stationary, there is no change between two consecutive images, so that it cannot be detected and the alarm is lost.

In order to improve this drawback, as shown in FIG. 4, the current image obtained by the image pickup means 1 and the image input means 2 and the previously stored reference image are stored in the current image memory 6 and the reference memory 7, respectively. The difference between the current image and the reference image is calculated by the difference means 8 and the binarization means 5 is obtained.
After the binarization by a predetermined threshold value by the, the abnormality determination means 9 determines that the number of changed pixels is abnormal, and
There is a method of outputting by the abnormal output means 10 (Japanese Patent Laid-Open No. 60-7593).

However, in such an apparatus, when a change in brightness due to a phenomenon that is not originally abnormal, such as a change in illuminance due to sunlight, occurs in the monitored area, the brightness of the current image changes, so the difference means 8 keeps the brightness constant. When the difference from the reference image is calculated, the abnormality determination unit 9 has a problem that the number of changed pixels becomes larger than the set number and is determined to be abnormal, resulting in a false alarm.

In order to improve such a problem, the present inventors have proposed a device as shown in FIG.
-326256). That is, the input image as shown in FIG. 6 (a) is subjected to edge extraction by the edge extraction unit 3 as shown in FIG. 6 (b), and an intruder or the like is detected based on the extracted edge information. Is.

[Problems to be Solved by the Invention]

However, in the above device, the edge is extracted by differentiating the image signal, and the edge portion is unlikely to change even when the optical conditions such as illumination change, so the false alarm is temporarily improved. However, for example, there is an object B in the monitoring area, there is a strong edge such as a shadow K due to sunlight, the position of the sun moves over time, and the edge due to this shadow is also as shown in FIG. 6 (c). When the edge image of FIG. 6 (b) is stored as a reference image when moving to, the edge image of FIG. 6 (c) is compared with the reference image and the edge image of FIG. As a result, there is a problem in that a false edge is generated due to the moved edge. Here, the dotted line indicates an edge disappeared by comparison.

The present invention has been made in view of the above points, and an object of the present invention is to provide a highly reliable image recognition sensor with reduced false alarms.

[Means for solving the problem]

The abnormality monitoring device of the present invention includes an image pickup unit for picking up a monitoring area, an image input unit for A / D converting an image signal output from the image pickup unit, and a current image for storing the A / D converted image signal. A memory; a reference image memory that stores in advance a reference image indicating a normal state of the monitoring area; a difference unit that compares the stored contents of the current image memory and the reference image memory to obtain a difference image; In an abnormality monitoring device including an abnormality determining unit that determines whether an abnormality is present in an image and an abnormality output unit that outputs the abnormality to the outside based on the determination result of the abnormality determining unit, the image output from the image input unit An edge extracting means for extracting an edge of an image from a signal, a binarizing means for binarizing an output of the edge extracting means, and an OR operation of an output of the binarizing means and an output of the reference image memory. Logical hand And the output from the OR means is stored in the reference image memory as a reference image, and the reference image and the current image are compared in place of the difference means. The present invention is characterized in that a logical operation means for extracting only newly appearing edges is added to the current image.

[Action]

In the abnormality monitoring device of the present invention, the temporal change of the background edge image is cumulatively stored as the reference image in the reference image memory by the OR operation means. The logical operation means compares the cumulatively stored reference image with the current image and extracts only the edge newly appearing in the current image.

〔Example〕

Hereinafter, the present invention will be described with reference to the drawings given as one embodiment.

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

Reference numeral 1 denotes a two-dimensional image pickup means such as a television camera, which picks up an image of the monitoring area and outputs an image signal. An image input unit 2 obtains a digital image signal by A / D converting the image signal. Reference numeral 4 denotes an edge extraction means, which two-dimensionally differentiates the image signal and extracts only the edge having a large luminance change. The edge extracting means 4 can be realized by a 3 × 3 filter such as a Sobel operator. The Sobel operator is a method that is often used when performing edge extraction, and the brightness at coordinates (x, y) is I
Assuming that (x, y), the brightness J (x, y) of the edge image is obtained by the following formula.

J (x, y) = | I (x-1, y-1) + I (x-1, y) + I
(X-1, y + 1) -I (x + 1, y-1) -I (x + 1, y)
−I (x + 1, y + 1) | + | I (x−1, y−1) + I (x, y
−) + I (x + 1, y−1) −I (x−1, y + 1) −I (x,
y + 1) -I (x + 1, y + 1) | 5 is a binarizing means, which binarizes the edge image extracted by the edge extracting means 4 by a preset threshold value, and the edge portion is "1", other portions. Is set to “0. 6 is a current image memory for storing the edge image of the current image.
Reference image memory 7 stores the edge image of the reference image indicating that the monitoring area is in a normal state. Reference numeral 11 is a logical operation means for performing a predetermined logical operation, which compares the edge image of the current image with the edge image of the reference image, and obtains only the edges that do not appear in the reference image in the current image. Reference numeral 9 is an abnormality determining means for determining whether or not there is an abnormality based on the output of the difference means 8. When the number of changed pixels in the difference image is larger than a predetermined number, it is determined that there is an abnormality. Reference numeral 10 denotes an abnormality output means, which outputs an abnormality to the outside by a display device or the like when the abnormality determination means 9 determines that there is an abnormality.

Reference numeral 12 is a logical sum operation means, which takes the logical sum of the output of the binarization means 5 and the output of the reference image memory 7 and stores the result as the reference image memory. That is, the reference image is stored by accumulating the temporal changes of the edge image generated by sunlight or the like.

Next, the operation of this embodiment will be described. This operation consists of a setting mode and a detection mode.

(Setting Mode) First, the background image when there is no intruder or the like is stored in the reference image memory 7 as a reference image. The edge image does not change much with changes in illuminance within the surveillance area,
Since the irradiation angle of sunlight from a window or the like changes over time, the edge also changes. The change amount of the edge caused by the sunlight is cumulatively stored in the reference image memory 7 via the logical sum calculation means 12. For example, as shown in FIG. 2A, it is assumed that the object B exists in the monitoring area and the shadow K of the object B exists due to the influence of sunlight. The edge image in this case is the second
It becomes like FIG. Due to the movement of the sun, the edge image of the shadow K moves as shown in FIG. 2 (c). The reference image is
As shown in FIG. 2D, the edge images shown in FIGS. 2A and 2B are accumulated.

(Detection mode) After preparing for the setting mode, enter the detection mode.
Now, assuming that the intruder S appears, the edge image of the current image is as shown in FIG. 2 (e) and is stored in the current image memory 6. An edge image as shown in FIG. 2 (d) is stored in the reference image memory 7, and the logical operation means 11 compares the reference image with the current image to perform a logical operation as shown in FIG. I do. That is, as shown in FIG. 2 (f), only the edges newly appearing in the current image are extracted. Here, the dotted line in FIG. 3 (f) indicates an edge disappeared by comparison. Edges that were in the reference image but are not the current image, or edges that are present in both the current image and the reference image are not extracted.

Therefore, in the present embodiment, even if the edge of the reference image moves due to a change in sunlight or the like, no false alarm is generated due to this.

The output of the logical operation means 11 is determined to be abnormal by the abnormality determining means 9 when the number of pixels of the edge by the intruder S has become a predetermined threshold abnormality, and the abnormality output means 10 informs the outside.

〔The invention's effect〕

As described above, according to the present invention, in the abnormality monitoring device that determines an abnormality by comparing the reference image and the current image using the image signal captured from the image pickup means, an edge that extracts an edge of the image from the image signal. The extraction means, the binarization means for binarizing the output of the edge extraction means, and the OR operation means for performing the OR operation of the output of the binarization means and the output of the reference image memory are added, and the logic is added. A logical operation in which the output from the summing means is stored as a reference image in the reference image memory, the reference image and the current image are compared, and only edges newly appearing in the current image that are not in the reference image are extracted. Since the means is added, it is possible to provide a highly reliable abnormality monitoring device with reduced false alarms.

[Brief description of drawings]

 FIG. 1 is a block diagram showing an embodiment of the present invention, FIG. 2 is a monitor screen diagram for explaining the operation of the above, FIG. 3 is a logic diagram for explaining the operation of the same, and FIG. Is a block diagram showing a conventional example, FIG. 5 is a block diagram showing another conventional example, and FIG. 6 is a monitoring screen diagram for explaining the operation of the same. 1 ... Imaging means, 2 ... Image input means, 4 ... Edge extraction means, 5 ... Binarization means, 6 ... Current image memory, 7 ... Reference image memory, 9 ... Abnormality determination means, 10 ... Abnormality Output means 11 ... Logical operation means, 12 ... Logical OR operation means

Claims (1)

(57) [Claims] 1. An image pickup means for picking up an image of a monitoring area, an image input means for A / D converting an image signal output from the image pickup means, and a current image memory for storing the A / D converted image signal. A reference image memory that stores in advance a reference image indicating the normal state of the monitoring area, a difference unit that compares the stored contents of the current image memory and the reference image memory to obtain a difference image, and an abnormality from the difference image. In an abnormality monitoring device including abnormality determination means for determining whether or not the abnormality is determined, and abnormality output means for outputting an abnormality to the outside according to the determination result of the abnormality determination means, an image is output from the image signal output from the image input means. Edge extraction means for extracting the edges of the, the binarization means for binarizing the output of the edge extraction means, and the OR operation means for performing the OR operation of the output of the binarization means and the output of the reference image memory. And add The output from the logical sum means is stored in the reference image memory as a reference image, and instead of the difference means, the reference image and the current image are compared, and the current image is newly added instead of the reference image. An abnormality monitoring device characterized in that a logical operation means for extracting only the appeared edges is added.