JPH0816234A - Device for detecting abnormality of plant - Google Patents

Abstract

PURPOSE:To surely separate and discriminate whether there is the intrusion of a foreign body such as a human the generation of smoke at the time of detecting abnormality and to attain instantaneous separation/discrimination without requiring timewise processing. CONSTITUTION:When a monitoring video is sent from a television camera 10 to a plant abnormality detector 1, an A/D conversion part 2 converts the video into a digital picture, and a difference processing part 3 finds out 3 difference picture between the digital picture and a reference picture stored in a reference picture storing memory 7 and stores the difference picture in a difference picture storing memory 8. A picture state judging part 5 judges whether a binarized area is larger than the difference picture or not, and when the area is larger, judges the existence of abnormality. A differential processing part 4 executes expansion processing after the binarization of the difference picture stored in the memory 8, executes the differential processing of the difference picture, stores the processed result in a differential picture storing memory 9, and extracts the differential picture of a changed part from the processed result. The judging part 5 judges the existence of a foreign body such as a human when the contour area of the differential picture is larger and judges the generation of smoke when the area is smaller.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a plant abnormality detecting apparatus by contour judgment for automatically detecting abnormality in a plant.

[0002]

2. Description of the Related Art In recent years, various devices have been proposed for automatically detecting an abnormality in a plant by image processing without the need for a person to visit the site or visually check a surveillance television. For example, when detecting an abnormality such as an oil leak, a steam leak, a human invasion, or smoke generation, a method of comparing a video in which no abnormality has occurred and a video in which abnormality has occurred and detecting a different portion is used. The image in which no abnormality has occurred is stored in advance as a reference image and is compared with the image captured at the time of inspection. A plurality of reference images are prepared according to the state of the plant, or are automatically updated by following changes in the state of the plant. The reference image and the images captured at the time of inspection emphasize the features of the image through filters and masks to make it easier to detect different areas, and several images to reduce unnecessary noise areas. They are averaging.

FIG. 11 shows an example of a plant abnormality detecting apparatus in the conventional system. In FIG. 11, reference numeral 10 denotes a plurality of television cameras or infrared cameras, the video signals of which are selected by the camera selector 11 and the plant abnormality detection device 2
Sent to 1. This plant abnormality detection device 21 has A
/ D conversion unit 22, difference calculation unit 23, image state determination unit 2
4, video storage memory 25, differential video storage memory 26
Is provided.

The plant abnormality detection device 21 converts the image selected by the camera selector 11 into a digital image by the A / D converter 22 and stores it in the image storage memory 25. The reference image is stored in advance in the image storage memory 25. The reference image stored in advance in the image storage memory 25 and the image captured at the time of inspection are compared in the difference calculation unit 23, and different portions are detected. The difference image obtained by the difference calculation unit 23 is stored in the difference image storage memory 26. The image state determination unit 24 includes a difference video storage memory 26.
Whether the status of the differential image stored in the memory is abnormal or not is determined, and when an abnormality is detected, an abnormal signal is output to the series control computer 12
Output to. When the abnormality signal is sent from the plant abnormality detection device 1, the sequence control computer 12 issues an alarm to notify the supervisor that the abnormality has occurred.

[0005]

As described above, in the conventional system, the image in the normal state in which no abnormality has occurred is stored as the reference image in the differential image storage memory 26, and is stored in the television camera 10. The difference image from the captured image is binarized, and whether or not an abnormality has occurred is determined by focusing on the area and the movement state of the center of gravity. For example, when smoke is generated in the monitoring area, the generated smoke portion appears as a difference in the difference image output from the difference calculation unit 23, so that it is possible to detect that an abnormality has occurred.

However, with this method, even if a foreign object such as a human invades, a difference appears in the difference image, so that it is difficult to detect only smoke separately. In order to eliminate this difficulty, the change in the difference over time is captured, and if there is a foreign object such as a human, it moves over time, but smoke is generated from one place, so there is no overall difference. It uses a method of making a separation judgment based on the state of change.

However, this method requires a considerable amount of time between the invasion of foreign matter by humans and the separation of smoke generation, and the misjudgment of smoke when the foreign matter stays in its place after invasion. And if the difference is small,
The drawback was that it was difficult to capture changes over time.

The present invention has been made in view of the above-mentioned circumstances, and when an abnormality is detected, it is possible to reliably determine the separation of a foreign substance such as a human being and the smoke generation, and it is possible to perform an immediate separation determination without the need for temporal processing. An object is to provide a plant abnormality detection device.

It is another object of the present invention to provide a plant abnormality detecting apparatus capable of correctly separating and discriminating the intrusion of foreign matter such as a human and the generation of smoke even when the difference between the reference image and the captured image is small. To do.

[0010]

SUMMARY OF THE INVENTION The present invention is an abnormality detection apparatus for capturing an image from a monitoring camera installed at an arbitrary monitoring place in a plant to detect the presence or absence of an abnormality in the plant, which is output from the monitoring camera. A / D conversion means for converting the image to be converted into a digital image, a difference processing means for obtaining a difference image between the image captured through this means and a reference image having no abnormality, and a difference obtained by this means It is characterized by further comprising a determining means for determining the presence or absence of abnormality by measuring the area of the contour included in the portion where the change of the image is seen.

Further, according to the present invention, the content of the contour area with respect to the area of the difference image is further measured with respect to the judgment result of the above-mentioned judgment means to separately judge the intrusion of foreign matter and the generation of smoke or the like at the minute change. It is characterized by

[0012]

The difference between the image features of smoke and foreign matter such as human beings is that there is no clear contour in smoke, but there is a clear contour in foreign matter such as humans. When the area of a large value in the differential image obtained by differentiating the difference image is defined as the contour area, the contour area is large for foreign matter such as humans and small for smoke, so
By measuring this area, it is possible to separate smoke and foreign matter such as humans.

However, in practice, even in a portion where there is little difference between the reference image and the captured image due to image distortion or the like, the differential value becomes large and appears as an area, resulting in erroneous detection. In order to exclude these inappropriate areas and obtain the correct contour area,
An image obtained by performing noise reduction and binarization and expansion processing from the difference image is used as a mask image, and only a portion passing through the mask image is used as an image for abnormality determination. The dilation process performed here is performed based on the principle that a region slightly larger than the target part may be used as the extraction region because the contour part of the target part mainly occurs at the boundary part of the difference image. There is. By doing so, only the contour portion of the target portion is obtained, and by measuring the area of the obtained contour portion, smoke and foreign matter such as a human can be correctly separated and detected.

If the intrusion of foreign matter is very small, the contour area obtained by the above method becomes small, and the separation determination may not be performed correctly. However, by calculating the content rate of the contour area with respect to the difference area obtained by binarizing the difference image, the content rate of the contour area is large when the intruder is a foreign substance such as a human, and the content of the contour area is included when the smoke is Since the rate becomes small, it becomes possible to determine the separation of smoke and foreign matter such as human beings when a minute change occurs.

[0015]

An embodiment of the present invention will be described below with reference to the drawings. FIG. 1 shows an embodiment in which an abnormality detection system is constructed using the plant abnormality detection device according to the present invention. Further, FIG. 2 is a flowchart showing the processing operation of the abnormality detection in FIG. In FIG. 1, 10
Is a plurality of TV cameras (or infrared cameras) for monitoring plant abnormalities, and its video output is the camera selector 11
And is sent to the runt abnormality detecting device 1. The plant abnormality detection device 1 includes an A / D conversion unit 2, a difference processing unit 3, a differentiation processing unit 4, an image state determination unit 5, a minute determination unit 6,
It comprises a reference image storage memory 7, a difference image storage memory 8 and a differential image storage memory 9. The reference image storage memory 7 stores a video to be monitored in advance in a state where no abnormality has occurred, for example, a reference video as shown in FIG. In FIG. 3, 13 is a device to be monitored.

A series control computer 12 is connected to the downstream side of the plant abnormality detection device 1 and issues an alarm when the plant abnormality detection device 1 outputs an abnormality signal. Next, the operation of the above embodiment will be described with reference to the flowchart of FIG.

An image of the monitoring area of the plant is taken by the television camera 10. This TV camera 1
The images captured by 0 are collected by the camera selector 11, one required image is selected and sent to the plant abnormality detection device 1 (step A1).

Now, assuming that an image containing a foreign matter 14 such as a human being is sent from the camera selector 11 to the plant abnormality detection apparatus 1 as shown in FIG.
It is converted into a digital image by the / D converter 2. The digitized monitoring target image is sent to the difference processing unit 3 and subjected to difference processing with the reference image shown in FIG. 3 stored in the reference image storage memory 7. As a result,
The difference image 15 of the foreign substance portion such as a person as shown in FIG. 3 is obtained and stored in the difference image storage memory 8 (step A
2).

Then, the image state judging section 5 judges whether or not the area after binarization is large in the difference image (step A3), and if the area after binarization is not large,
It is determined that there is no change (normal) (step A4), and the determination result is output to the minute determination unit 6.

Further, in step A3, if the area after binarization is large, it is determined that there is an abnormality, and the differential processing section 4 stores the difference image storage memory 8 as shown in FIG. After removing noise from the difference image, 2
After the binarization, the dilation process is performed to create the binarized dilated image 16 of only the changed portion as shown in FIG. 6 (step A6).
Further, the differential image 17 as shown in FIG. 7, which is obtained by performing the differential processing on the portion where the value appears in the differential image, is stored in the differential image storage memory 9 (step A7). Furthermore, step A
A differential image of the changed portion is extracted from the processing results of 6 and A7 (step A8). That is, an image that has been subjected to noise removal and binarization and expansion processing in order to correctly obtain the contour area is used as a mask image, and only a portion that has passed through the mask image is used as an image for abnormality determination.

Here, when a foreign substance such as a human invades the monitoring area, the contour clearly appears as in the differential image 17 shown in FIG. 7, and the contour area becomes large. However, in the case of the image in which the smoke 18 is generated as shown in FIG. 8, the difference image 19 of the smoke portion appears as shown in FIG. 9, but since there is no clear contour, the difference is differentiated as shown in FIG. Image 2
No contour appears at 0, and the contour area becomes small.

Therefore, the image state judging unit 5 checks whether or not the contour area in the binary image of the differential image is large (step A9), and if the contour area is large, it is judged as a foreign substance such as a human ( In step A10), if the contour area is small, it is tentatively determined that smoke is generated (step A12), and the determination result is output to the minute determination unit 6.

In this minute determination section 6, step A13,
The contour content rate determination module composed of A14 determines the contour content rate. However, if the determination from the image state determination unit 5 is that there is no change or a foreign substance such as a human being, the determination result is directly sent to the series control computer 12. Output (Step A
5, A11). However, if it is determined to be smoke, the contour portion content rate is calculated from the difference area and the differential area (step A
13) It is checked whether or not the contour portion content rate is larger than a preset value (step A14). If the contour portion content rate is large, it is determined that the foreign matter is a human or the like, and the corrected determination result is sequence controlled. Output to computer 12 (step A1)
1). Further, the minute determination unit 6 determines that smoke is generated when the above-mentioned content ratio of the contour portion is small (step A
15), and outputs the determination result to the series control computer 12 (step A16).

If the intrusion of foreign matter is very small, the contour area obtained by the above method becomes small, and the step A
In some cases, the separation determination in 9 cannot be performed accurately. Therefore, if it is determined in step A9 that the contour area is small, it is temporarily determined that smoke is generated, and then
The minute determination unit 6 obtains the content rate of the contour area with respect to the difference area obtained by binarizing the difference image. If the invaded matter is a foreign substance such as a human being, the content rate of the contour area is large, and if it is smoke, the content rate of the contour area is small. It will be possible.

[0025]

As described above in detail, according to the present invention, the abnormality is judged by grasping the state of the contour of the portion where the monitor image has changed with respect to the reference image. The smoke generation can be reliably separated, and the immediate separation determination can be performed without the need for time processing.
Further, since it does not depend on the change over time, it is possible to always make a correct judgment even if a foreign object such as an intruding person stays on the spot after the intrusion. Further, even when the difference between the reference image and the captured image is small, by determining the content rate of the contour area with respect to the difference area, it is possible to correctly separate and discriminate between foreign matter such as human being and smoke generation.

[Brief description of drawings]

FIG. 1 is a block diagram showing the configuration of a plant abnormality detection apparatus according to an embodiment of the present invention.

FIG. 2 is a flowchart of abnormality detection in the embodiment.

FIG. 3 is a conceptual diagram showing an image serving as a criterion for determination.

FIG. 4 is a conceptual diagram showing an image when a foreign matter such as a human invades.

FIG. 5 is a conceptual diagram showing a difference image when a foreign matter such as a human invades.

FIG. 6 is a conceptual diagram showing an image obtained by binarizing and expanding a difference image when a foreign matter such as a human intrudes.

FIG. 7 is a conceptual diagram showing a contour image when a foreign matter such as a human invades.

FIG. 8 is a conceptual diagram showing an image when smoke is generated.

FIG. 9 is a conceptual diagram showing a differential image when smoke is generated.

FIG. 10 is a conceptual diagram showing an outline image when smoke is generated.

FIG. 11 is a block diagram showing the configuration of a plant abnormality detection device using a conventional method.

[Explanation of symbols]

 1 Plant Abnormality Detection Device 2 A / D Conversion Section 3 Difference Processing Section 4 Differentiation Processing Section 5 Image State Determination Section 6 Minute Determination Section 7 Reference Image Storage Memory 8 Difference Image Storage Memory 9 Differential Image Storage Memory 10 Television Camera or Infrared camera 11 Camera selector 12 Sequence control computer 13 Monitoring target device 14 Foreign body such as human being 15 Difference image of foreign body portion such as human 16 Image obtained by binarizing and expanding difference image of foreign body portion such as human being 17 Difference image of foreign body portion such as human being Differentiated image 18 Smoke 19 Difference image of smoke part 20 Image differentiated difference image of smoke part

Continuation of front page (51) Int.Cl. ⁶ Identification number Office reference number FI Technical display location G08B 25/00 510 M 0803-2E H04N 7/18 D

Claims (2)

[Claims] 1. An abnormality detection device for detecting the presence or absence of abnormality in a plant by capturing an image from a monitoring camera installed at an arbitrary monitoring location in a plant, and converting the image output from the monitoring camera into a digital image. A / D conversion means, a difference processing means for obtaining a difference image between the image captured through this means and a reference image having no abnormality, and a portion where a change in the difference image obtained by this means is seen. A plant abnormality detection apparatus comprising: a determination unit that determines the presence or absence of abnormality by measuring the area of the included contour. 2. The intrusion of foreign matter and the occurrence of smoke or the like at the time of a minute change are separately determined by measuring the content rate of the contour area with respect to the area of the difference image with respect to the determination result of the determination means. The plant abnormality detection device according to claim 1.