JPH1097687A - Monitoring device - Google Patents

Abstract

PROBLEM TO BE SOLVED: To protect a monitoring device from discriminating the variation of a monitoring picture as abnormality when the change is caused by the movement of a registered picture by registering pictures which are predicted to move in a background picture. SOLUTION: A background picture from a camera 1 is converted into digital data by means of an A/D converter 2 and written in a background picture memory 10 and a picture memory 3 and pictures which are predicted to move are set by means of a setting means 6, extracted by means of a picture extracting section 7, and registered in a registered-picture memory 8. The picture supplied from the A/D converter 2 from one minute to the next are written in a monitored-picture memory 11 and the correlation between the pictures from the memories 10 and 11 are calculated by means of a correlation calculating section 12. When a difference discriminating section 14 discriminates a difference between the pictures, a correlation calculating section 16 calculates the correlations between the pictures from the memories 11 and 8 and an abnormality discriminating section 17 discriminates the presence of abnormality, and then, a clock section 18 counts time. When the difference continuously exist even after a required period of time counted by means of a time discriminating section 19 has elapsed, the picture which is stored in an abnormal-picture memory after the picture is transferred to the memory from the memory 11 is read out from the memory and outputted from an output section 21.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a monitoring apparatus, and more particularly to a monitoring apparatus for registering a background image and an image that may move, so that when the registered image moves, it is not determined that the image is abnormal.

[0002]

2. Description of the Related Art In a device for detecting a change in a monitoring image taken by a television camera (hereinafter abbreviated as a camera) and notifying an abnormality, a background image when there is no abnormality is stored, and this image is sequentially output from the camera. This is compared with the monitoring image to be detected, and when a required difference is determined, it is determined that an abnormality has occurred. The monitoring image is also used for monitoring theft of an article or monitoring the leaving of a suspicious object. However, there is a problem that it is determined that an abnormality has occurred even when a part of an image stored as a background image has moved, and in a place where people frequently enter and leave, the movement of a person in a normal range is determined to be abnormal. There is a problem of doing it.

[0003]

SUMMARY OF THE INVENTION In view of the above, in the present invention, a part of the background image moves, that is, it should move like a fixture without making the movement of a person in a normal range abnormal. It is to judge that an abnormality has occurred when something missing from the screen disappears, or when an article that is not in the background image is placed.

[0004]

SUMMARY OF THE INVENTION In order to solve the above-mentioned problems, the present invention monitors an abnormality based on a difference between a background image when there is no abnormality from a camera and a monitoring image sequentially output from the camera. A background image memory for storing the background image, a setting unit that sets an image predicted to move in the background image, an image extraction unit that extracts an image set by the setting unit from the background image, A registered image memory for storing an image from the image extracting unit, a monitored image memory for storing the monitoring image, and a first correlation operation for calculating a correlation between the image read from the background image memory and the image read from the monitoring image memory. Unit, a difference discriminating unit for discriminating the presence or absence of an image difference based on data from the first correlation calculating unit, and an image from the registered image memory and a monitor image based on the discrimination of the image difference by the difference discriminating unit. A second correlation operation unit for calculating the correlation of the image from the moly; an abnormality determination unit for determining an abnormality based on data from the second correlation operation unit; and a clock for starting timekeeping when the abnormality determination unit determines that there is an abnormality. Unit, and a time determination unit that outputs a signal when a required time has elapsed from the start of timekeeping of the clock unit, and when the signal is output from the time determination unit, the difference determination unit continued to determine the difference. A monitoring device that outputs an abnormal image in a case is provided.

[0005]

DESCRIPTION OF THE PREFERRED EMBODIMENTS In a monitoring apparatus according to the present invention, a background image when there is no abnormality from a camera is stored in a background image memory, and an image whose motion is predicted by the background image is set by a setting unit. The set image is extracted from the background image by the image extracting unit, and the extracted image is stored in the registered image memory. An image sequentially output from the camera is stored in a monitoring image memory, and a correlation between the image read from the background image memory and the image read from the monitoring image memory is calculated by a first correlation calculating unit. The data is input to the difference discriminating unit, and is compared with the threshold value read from the storage unit. If the difference is smaller than the threshold value, the difference is discriminated. In the discrimination of the difference, the image from the registered image memory and the image from the monitoring image memory are determined. Is calculated by the second correlation calculation unit, the data from the second correlation calculation unit is input to the abnormality determination unit, and is compared with the threshold value from the storage unit. When the time determination unit determines that the required time has elapsed, if the difference determination unit continues to determine the difference, the image transferred from the monitoring image memory to the abnormal image memory is read and output.

[0006]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS An embodiment of a monitoring device according to the present invention will be described below in detail with reference to the drawings. FIG. 1 is a block diagram of a main part of an embodiment of a monitoring device according to the present invention. In the figure, 1
Denotes a camera, which captures an image of a monitoring area with a fixed viewing angle. 2
Denotes an A / D converter, which converts a video signal from the camera 1 into a digital signal. Reference numeral 3 denotes an image memory, which is an A / D converter 2
The image (image data) when no abnormality is detected is stored. Reference numeral 4 denotes a D / A converter for converting image data read from the image memory 3 into an analog video signal. Reference numeral 5 denotes an image display unit that displays an image on the screen based on the video signal from the D / A conversion unit 4. Reference numeral 6 denotes a setting unit that predicts a moving object in the image on the image display unit 5 and specifies and sets the coordinates of a plurality of external shapes with a pointing device. Reference numeral 7 denotes an image extracting unit that extracts the image set by the setting unit 6 from the image read from the image memory 3. Reference numeral 8 denotes a registered image memory for storing the image extracted by the image extracting unit 7. 9 is a control unit,
It controls the image extraction unit 7 and the registered image memory 8.

Reference numeral 10 denotes a background image memory which inputs and stores a background image from the camera 1 through the A / D converter 2 when there is no abnormality. Reference numeral 11 denotes a surveillance image memory, which inputs and stores images that are sequentially output from the camera 1 via the A / D converter 2. Reference numeral 12 denotes a correlation operation unit that calculates a correlation between the image read from the background image memory 10 and the image read from the monitoring image memory 11 (calculates similarity, and outputs a large signal when there is a correlation). A storage unit 13 stores a threshold A (first threshold) as a criterion for determining the difference between the background image and the monitoring image, and a threshold B (second threshold) as a criterion for determining the difference between the registered image and the monitoring image. I do. Reference numeral 14 denotes a difference discriminating unit which compares data from the correlation calculating unit 12 with a threshold A read from the storage unit 13 and
If smaller, it is determined that there is an image difference. Reference numeral 15 denotes an area setting unit which sets an image area for which the difference determination unit 14 has determined a difference. Reference numeral 16 denotes a correlation calculation unit which calculates a correlation between the image from the registered image memory 8 and the image of the area set by the area setting unit 15 from the monitoring image memory 11 when the difference determination unit 14 determines that there is an image difference. . Reference numeral 17 denotes an abnormality determination unit which compares data from the correlation operation unit 16 with a threshold value B read from the storage unit 13, and determines that there is an abnormality when the data is smaller than the threshold value B. Reference numeral 18 denotes a clock unit, which starts timing when the abnormality determination unit 16 determines that there is an abnormality. Reference numeral 19 denotes a time determination unit which outputs a signal when a required time has elapsed since the clock unit 18 started counting time. Reference numeral 20 denotes an abnormal image memory, which stores an image read from the monitoring image memory 11 and transferred. Reference numeral 21 denotes an output unit for reading and outputting an image in the abnormal image memory 20.
A control unit 22 controls each unit.

Next, the operation of the monitoring apparatus according to the present invention will be described with reference to FIG.
3 will be described with reference to the image diagram shown in FIG. First, prior to monitoring, the video signal of the background image (a in FIG. 2) from the camera 1 is converted into a digital signal by the A / D converter 2 (step 1, hereinafter abbreviated as ST1) and stored in the background image memory 10. Store one screen (ST2),
At the same time, this image is stored in the image memory 3. The image (image data) stored in the image memory 3 is read out, and D / A
The signal is converted into an analog signal by the conversion unit 4, input to the image display unit 5, and displayed on a screen. The camera 1 is installed in the cash corner of the financial institution in the example of the figure, 31 is equipment, 32 is an article which is to be moved by a clerk and should disappear from the screen, and 33 is a person. Therefore, the setting unit 6 uses a pointing device (such as a mouse) on the screen of the image display unit 5 to
The coordinates of the outer shape of 32 and the coordinates of the outer shape of the upper half of the person 33 (portion surrounded by the dotted line) are designated (set) (ST3). With this setting, the article 32 and the person 33 are extracted from the background image (a) read from the image memory 3 by the image extracting unit 7 via the control unit 9 and input to the registered image memory 8, and FIG.
The image of the article 32 is registered (stored) at the position corresponding to the article 32 in (A), and the image of the person 33 is registered (ST4).

Thereafter, the video signal sequentially output from the camera 1 is converted into a digital signal by the A / D converter 2 and stored in the monitoring image memory 11 (ST5). Then, images are read out from the background image memory 10 and the monitoring image memory 11, and these are input to the correlation operation unit 12, and the correlation is calculated (ST6). In calculating the correlation, for example, the two images are divided into a plurality of sections each having the same size, the correlation (similarity) is calculated in each section, and a signal corresponding to the magnitude of the correlation is output. The data from the correlation calculation unit 12 is input to the difference discrimination unit 14 and is stored in the storage unit via the control unit 22.
13 is compared with the threshold value A read out. If the monitoring image at this time is an image in which the article 32 has disappeared as shown in FIG. 2B, the data from the correlation calculation unit 12 is smaller than the threshold value A, and the difference determination unit 14 determines whether there is an image difference. (ST7: Yes). In this determination, the image area (the part of the article 32) determined to have a difference by the area setting unit 15 via the control unit 22 is set.

Based on the determination that there is an image difference, the control unit 22
The area set by the area setting unit 15 via the
The image of (32) is cut out from the image in the monitoring image memory 11 and input to the correlation calculation unit 16 together with the image from the registered image memory 8 to calculate the correlation (ST8). This data is input to the abnormality determination unit 17 Is compared with the threshold value B read from the storage unit 13 via the control unit 22. In this comparison, the surveillance image (Fig. 2b)
Since the article 32 that has disappeared in step S1 is one of the images registered in the registered image memory 8, the data (correlation value) from the correlation operation unit 16 is larger than the threshold value B (ST9: Yes). Is determined to be normal (ST10).

When the article 31 (equipment, etc.) disappears or a new article 34 is placed as shown in FIG. It is determined whether there is a difference (ST7: Yes), and the image of the area (the part 31 and the article 34 in FIG. 2C) set by the area setting part 15 via the control part 22 is monitored via the control part 22. Memory 11
From the registered image, and input the image together with the image from the registered image memory 8 to the correlation operation unit 16 to calculate the correlation (ST ST).
8). This data is input to the abnormality determining unit 17 and is compared with the threshold value B read from the storage unit 13. Since it is smaller than the threshold value B (no correlation) (ST9: No), it is determined that there is an abnormality (ST11).
In this determination, the clock unit 18 starts counting time (ST12), and the time determination unit 19 determines that the required time has elapsed while the difference determination unit 14 continues to determine whether there is an image difference (ST13: Yes). (ST14:
Yes), abnormal image memory transferred from monitoring image memory 11
20 is read out and output from the output unit 21 (ST
15) Displayed on a monitor screen or the like (not shown), or output an alarm for occurrence of abnormality at the same time.

Further, as shown in FIG.
When 35 appears, since the data calculated by the correlation calculator 12 is smaller than the threshold value A, the difference discriminator 14 discriminates that there is an image difference (ST7: Yes), and the image from the monitor image memory 11 and the registered image memory 8 The data calculated by the correlation calculator 16 with the image is larger than the threshold value B (ST9: Yes). That is, the person 35 is determined to be one of the registered images, and the movement of this image is not abnormal, but is determined to be normal behavior, and it is determined that there is no abnormality (ST10). This determination is made, for example, by calculating the ratio between the head size and the shoulder width of the person 35 and the person 33, comparing the two, and if the required match rate (set lower than the match rate of the article 31, etc.) is equal to or greater than The person 35 is determined to be one of the registered images.

[0013]

As described above, according to the monitoring apparatus of the present invention, an image predicted to move in the background image is registered in advance, so that even if this part (object) moves, abnormal It does not incorrectly determine the occurrence, and also does not determine that a person has appeared, as well as an abnormality, and determines that an abnormality has occurred if equipment is removed or a suspicious object is placed. Therefore, it is possible to improve the reliability of the monitoring device that monitors only what is needed.

[Brief description of the drawings]

FIG. 1 is a main part block diagram of an embodiment of a monitoring device according to the present invention.

FIG. 2 is an example of an image diagram for explaining the operation of the monitoring device according to the present invention.

FIG. 3 is a flowchart for explaining the operation of the monitoring device according to the present invention.

[Explanation of symbols]

 DESCRIPTION OF SYMBOLS 1 Television camera 2 A / D conversion part 3 Image memory 6 Setting part 7 Image extraction part 8 Registration image memory 9, 22 Control part 10 Background image memory 11 Surveillance image memory 12, 16 Correlation calculation part 13 Storage part 14 Difference discrimination part 15 Area setting section 17 Abnormality determination section 18 Clock section 19 Time determination section 20 Abnormal image memory 21 Output section

Claims (6)

[Claims] 1. A background image memory for monitoring an abnormality based on a difference between a background image when there is no abnormality from a television camera and a monitoring image sequentially output from the television camera, wherein the background image is stored. A setting unit for setting an image predicted to move in the background image,
An image extraction unit that extracts an image set by the setting unit from the background image, a registered image memory that stores an image from the image extraction unit, a monitoring image memory that stores the monitoring image,
A first correlation calculator for calculating a correlation between an image read from the background image memory and an image read from the monitor image memory; a difference determiner for determining whether there is an image difference based on data from the first correlation calculator; A second correlation calculating unit for calculating a correlation between the image from the registered image memory and the image from the monitoring image memory by determining whether there is an image difference by the difference determining unit, and determining abnormality based on data from the second correlation calculating unit A monitoring device configured to output an image read from the monitoring image memory based on the determination of the abnormality by the abnormality determining unit. 2. A coordinate setting function using a pointing device or the like is provided in the setting unit, an image whose motion is predicted in a background image is set by designating coordinates of a plurality of outer shapes, and the background image is set via the image extracting unit. The monitoring device according to claim 1, wherein the set image is extracted and stored in the registered image memory. 3. A storage unit for storing a first threshold value for a reference between the background image and the monitoring image and a second threshold value for a reference between the monitoring image and the registered image, wherein the difference determining unit includes: First
When the data from the correlation operation unit is smaller than the first threshold value from the storage unit, it is determined that there is an image difference, and in the determination that there is an image difference, the data from the second correlation operation unit is transferred from the storage unit by the abnormality determination unit. The monitoring device according to claim 1, wherein the abnormality is determined when the value is smaller than the second threshold value. 4. An image processing apparatus according to claim 1, further comprising an area setting section for setting an image area in which said difference is determined by said difference determining section, wherein said second correlation calculating section sets an image of an area set by said area setting section from said monitoring image memory. 4. The monitoring device according to claim 1, wherein a correlation between the registered image and an image from the registered image memory is calculated. 5. A timer for starting timekeeping when the abnormality determination unit determines that there is an abnormality, and a time determination unit for outputting a signal when a required time has elapsed from the start of timekeeping of the clock unit, wherein the difference determination is performed. 5. The monitoring device according to claim 1, wherein, when a signal is output from the time determination unit while the determination of the difference is continued, the image read out from the monitoring image memory is output. apparatus. 6. An abnormal image memory for storing an abnormal image,
An output unit for outputting an image read from the abnormal image memory, transferring and storing the image read from the monitoring image memory to the abnormal image memory, reading the image, and outputting the read image via the output unit. The monitoring device according to claim 3, 4 or 5.