JPH0765152A - Device and method for monitoring - Google Patents

Abstract

PURPOSE:To monitor a field instrument through a photography device and an image processor by detecting the position of a painter in a measuring instrument display screen image, comparing the detected painter position with the painter position in a normal state, and judging and displaying whether or not there is abnormality. CONSTITUTION:An images of a circular instrument 1 or a water gauge which is inputted through the photography device 3 such as a camera is sent to the image processor 4. The image processor 4 binarizes the image. Then a body in the image is labeled so as to find, for example, a circular instrument board in the image, and a circular shape coefficient is measured as to the labeled body to specify the circular instrument board. According to a grightness distribution, the angles of the painter in the top-bottom and right-left directions on the screen are found. The image processor 4 compares the value that the instrument indicates with an angle range of normal values and sends the value that the instrument indicates and a normal/abnormal signal to an alarm board 5.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a monitoring device and a monitoring method, and more particularly to a display value of a measuring instrument for measuring and displaying pressure, temperature, flow rate, concentration, pH value, etc. in a device such as a thermal power plant. The present invention relates to a monitoring device and a monitoring method for performing read monitoring with a target of abnormality monitoring.

[0002]

2. Description of the Related Art In plants such as thermal power plants, there is an increasing demand for automation of equipment monitoring due to demands for longer life of equipment and labor saving of patrols. Therefore, important equipment (for example, the fuel supply device in the boiler, the burner section, the main valves, etc.) is monitored using a surveillance TV camera,
The method of intensive monitoring on a monitor TV is widely used. Further, with the recent progress of image processing technology, FA (Factory) is often used to determine whether or not there is an abnormality by performing image processing on an image of a surveillance television camera.
It is increasing as a ripple from the field of automation.

However, the television camera has a limited monitoring area and cannot cover the entire boiler plant. Actually, a patrol officer visually inspects each part of the plant 6 to 7 times a day or a visual inspection by hearing. . This inspection inspection work requires a great deal of skill to detect anomalies, which is a considerable burden. In addition, there is a human limitation due to the centralization of monitoring or labor saving as a need of a power plant, and at present, a robot for patrol and inspect the site on behalf of patrol personnel is being developed.

When analyzing the function of human patrol,
(1) situation grasp, (2) movement, (3) situation judgment (normal or abnormal), (4) processing function, system function corresponding to this, (1) sensor technology, (2) movement Functions, (3) computer technology, and (4) driving guidance display.

When these systems are embodied and a robot inspection system is realized, the outline is as shown in FIG. 2. The three parts are as follows: (1) Surveillance robot 1, (2) ) An abnormality determination device 4, which processes information from the monitoring robot and determines whether there is an abnormality,
(3) Man-machine interface 5 that presents a raw image and a live sound from the scene and presents the details of the abnormality if the abnormality determination device determines that there is an abnormality.

The monitoring robot 1 includes an ITV camera 2 which is equivalent to the human senses of sight, hearing, smell and touch.
It is equipped with a microphone 3, a gas sensor and a non-contact type thermometer. These sensors are used to move around the site and collect information. This information is processed by an abnormality determination device such as image processing for image data and acoustic processing for acoustic data to determine the presence or absence of abnormality.

Here, an example of detecting an abnormality will be described. The items to be detected for anomalies vary depending on the monitoring target and the monitoring location, but in the case of a boiler plant, oil leaks and steam leaks are the main focus. These leaks lead to serious accidents if left unchecked, and early detection is required. The oil leak is accompanied by only the image change, and the image processing focusing only on the image change is performed to extract the abnormality. Hereinafter, the abnormality detection image processing procedure that is currently performed will be described by taking oil leak detection as an example.

An image processing flowchart is shown in FIG. The robot stores a normal state of the target device or place as an image in the memory 1 as a reference screen. Speaking of a boiler, it is an image of a place where an oil leak is likely to occur (a flange portion of an oil pipe, an oil pan in front of a burner, etc.) in a normal state. The binarization processing shown in FIG. 3 is to capture an image on a pixel-by-pixel basis.
This is a process of making a low one black, and is generally performed as a pre-process of image processing. In FIG. 3, when the robot makes a patrol and stops again at the place where the reference image is stored, the processing shown in the flowchart in the figure is performed. That is, the same target captured as the reference image is captured in the memory 2 as a target image, the image and the image are subtracted, the change amount (difference image) is extracted, and if this change amount is larger than a certain threshold value, the pixel is selected. Binarization processing is performed so that white and small ones become black. Here, for example, only the white ones are subjected to area calculation to determine whether or not they are abnormal. The reason for this area calculation is as follows.

In image-to-image matching, in general, even if images captured at the same location by a fixed camera are displaced by about one pixel. This is because a minute change, such as a vibration of the target device, which a person does not recognize. In addition, in the current surveillance system, the images are obtained by using a moving camera by a robot instead of a fixed camera, and positioning reproducibility is generally limited.
In order to prevent misjudgment of the amount of change due to this deviation, a threshold value is provided as a criterion. This standard is currently set at 3%. That is, the area is calculated, and if the change is within 3%, for example, it is considered as an image change due to the deviation, it is determined that there is no abnormality, and if it is 3% or more, it is determined that there is abnormality.

FIG. 4 shows an actual example in which abnormality processing is performed. This is an example of the case where the flange portion of the oil pipe is inspected. In the figure, the reference image is a normal image, and the target image is an image when oil leaks from the flange. This 2
By performing noise removal and binarization processing on one image, an image in which only leaked oil is detected can be obtained. This type of abnormality detection method is common, and is used as a product in abnormality determination logic such as a digital image sensor for unattended monitoring.

The example of abnormality detection determination using the configuration of the robot monitoring system in the plant and the image processing method has been described above.

[0012]

With the above-mentioned conventional technique, the leak of oil, water vapor, pulverized coal, etc. in the plant can be detected, so that the burden of the monitoring work on the patrol personnel in the plant is greatly reduced. Usually, patrols detect the above leaks and pressure gauges that have not been centrally monitored yet.
Field instruments such as thermometers and flow meters are also checked in parallel. It is desired to further promote the unmanned operation of the site by monitoring this field instrument with a monitoring device. Also, in the event of a large or small accident at the plant, it is essential to analyze the accident and clarify the cause of the accident whether or not the measured value at the time of the accident of the field instrument that was not incorporated in the central monitoring device was normal. There is an increasing need for monitoring these field instruments.

An object of the present invention is to solve the above-mentioned problems of the prior art and to provide a monitoring device and a monitoring method for monitoring an on-site instrument attached to a plant device or the like with a photographing device and an image processing device. .

[0014]

In order to achieve the above object, the invention claimed in the present application is as follows. (1) An image capturing device that captures an image of a monitored object and inputs the image, an image processing determination device that determines whether the device body including the monitored object is normal or abnormal based on the input image, and the determination of the device In a monitoring device provided with a device for displaying a result, a photographing means for photographing a region including a display surface of a weighing machine provided in a plant device and inputting the image, and a means for binarizing the inputted image A means for extracting a meter display surface image based on the feature amount of the meter display surface shape from the image after the above processing, the information of the extracted meter display surface image and the original image before the binarization processing. Based on the above, a means for detecting the pointer position on the image on the display surface of the meter in the original image and a means for comparing the detected pointer position and the normal pointer position to determine whether there is an abnormality and displaying it. A monitoring device provided.

(2) A step of photographing the monitored object and inputting an image thereof, a step of judging whether or not the information indicated by the monitored object is normal based on the input image, and a step of displaying the judgment result. In the monitoring method having, a step of photographing an area including a display surface of a weighing instrument attached to a plant apparatus and inputting the image, and binarizing the input image and the binarized image. A step of removing more noise and labeling for each shape object included in the image after the above processing, and for each labeled shape object, it is determined whether or not it corresponds to the display surface shape of the target weighing meter. Then, the process of extracting the image of the meter display surface from the image after the binarization process and the area of the meter display surface from the image before the binarization process based on the information of the extracted image of the meter screen. Decision process and decision For all of these areas, the step of obtaining the average luminance for each small area in the area, the step of detecting the position of the indicating means of the probing meter based on the average luminance, the position of the detected indicating means and the normal state And a step of determining whether there is an abnormality on the basis of the pointed position and displaying it.

(3) Monitoring provided with a photographing device for photographing the monitored object and inputting the image, and an image processing judging device for judging whether the information indicated by the monitored object is normal or not based on the inputted image. In the device, means for photographing an area including a display surface of a weighing instrument provided in the plant device and inputting the image, and a weighing meter from the input image based on the geometrical characteristic amount of the displaying surface of the weighing instrument. Means for obtaining the display surface area, means for obtaining the luminance distribution of the image within the obtained area, means for obtaining the position of the indicating means of the measuring instrument based on the luminance distribution, and of the measuring instrument based on the position of the indicating means. And a means for judging whether or not the information shown is within a normal range.

(4) A step of photographing the monitored object and inputting an image thereof, a step of judging whether or not the information indicated by the monitored object is normal based on the input image, and a step of outputting the judgment result. In a monitoring method including: a step of photographing an area including a display surface of a weighing instrument attached to a plant device and inputting the image; a step of binarizing the input image; A step of determining a search range for the weighing meter display surface in the input original image based on the image obtained, and a step of dividing the search range into a plurality of small areas to obtain an average luminance for each small area. A step of obtaining the position of the indicating means of the meter based on the obtained average luminance, a step of judging whether or not the information indicated by the measure meter is normal based on the position information of the indicating means, and outputting the judgment result. With process Monitoring method according to claim.

[0018]

The position of the indicator needle of a target image that can be extracted by image processing, for example, in a circular or band-shaped measuring meter, is determined from the marks attached to each measuring meter and the relative angle from the horizontal plane. Since it is possible to calculate the value indicated by each individual prometer, it is possible to confirm with sufficient accuracy the value of the instrument currently confirmed by inspection.

[0019]

FIG. 1 shows the configuration of the entire system. The images of the circular gauge 1 and the water level gauge 2 taken from the photographing device 3 such as a camera are sent to the image information processing device 4 and a device similar thereto, and subjected to noise removal such as smoothing and other preprocessing. By this preprocessing, the image of the needle of the instrument and its surroundings necessary for reading the value indicated by the instrument is extracted. After that, the image processing apparatus calculates the value indicated by the meter from the relative position between the position of the pointer of the meter and the mark attached to the meter for reading the pointer.

As a concrete example, FIG. 5, FIG. 6 and FIG. 7 show the steps of the image reading process of this circular instrument panel, and similarly, the image processing steps of reading the height of the float of the water level gauge. 8, FIG. 9 and FIG. First, FIG. 5, FIG.
The first embodiment will be described in detail with reference to FIG. The image of the monitoring target device captured by the monitoring camera is captured by the image processing device (51).

This image is immediately copied (52a),
Predetermined initial threshold value (th) (52c)
Is subjected to binarization processing (52b), and contraction / expansion processing is performed (53) to clarify the boundary line and to remove noise. Here, the contraction processing is, as shown in FIG. 11, for a white pixel in a binarized screen,
When all eight directions around the pixel are not white, it means to convert the pixel to black, and the expansion processing is performed on a black pixel on a binarized screen as shown in FIG.
If any of the pixels in eight directions around the pixel has white, it means converting the pixel to white, and by performing these processes, a value smaller than a certain size (noise).
Disappears, and larger ones remain. Then, in order to find the circular instrument panel in this image, labeling is performed on the objects in the image (54), and among many labeled objects, a circular-like object is found. At this time, the circular shape factor of each labeled object is measured (55a).

The circular shape factor is a factor of 1 if the object in the image is a perfect circle and approaches 0 each time the object moves away from it. Here, the circular shape factor is

[0023]

[Equation 1]

It means the value obtained in. This value is, for example, 0.
If it is larger than 9, the preprocessing is finished with this, and the binarized image is sent to the next step (55b). If this value is 0.9 or less and the labeling number limit of the image processing apparatus is not exceeded, the illuminance of the captured image is modulated, the image is captured again, and the processing up to this point is repeated (55c). .

If this value is 0.9 or less and exceeds the labeling number limit of the image processing apparatus, the binarization threshold of the captured image is changed to perform binarization, and the subsequent processing is performed. Repeat (55c). The contents of the flow chart described so far are for assuming that the most circular object in the captured image is the circular instrument panel and searching for that object (55d). Therefore, when capturing a circular instrument panel with a camera, it is necessary to select the camera position and angle for that purpose.

The image which has been processed up to this point shows an image in which the circular instrument panel is white and the pointer is black (FIG. 7). From this image, check what angle the pointer is making with respect to the vertical and horizontal directions of the screen. This process will be described below with reference to FIGS. First,
The center of gravity is searched for this binarized image, and its coordinates are stored as (x, y) (56a).

The distance from the (x, y) to the portion of the closest circumference is obtained, and the distance is stored as d (56
b). After obtaining these values, the following process is performed on the original screen of the image processed up to this point (the screen before binarization). Based on the values of (x, y) and d obtained earlier, a horizontally long window (processing area) including the position of the center of gravity is set (57a).

After that, while the original image is rotated by a small angle about the center of gravity of the circular instrument panel that substantially coincides with the axis portion of the pointer (57b), the average brightness of the processing area (window) is measured (57c). ). Generally, the dial of a circular instrument panel is white, and the pointer is colored red or black so that it has a good contrast. Therefore, if the average brightness is measured in this way, the average brightness of the processing region becomes lower than that of the part where the needles colored red and black are not (white instrument panel). The angle formed by the pointer of the instrument panel with the horizontal or vertical direction can be obtained from the rotation angle at which the average brightness is the lowest.

The above processing is performed for one rotation (360 °) (57d), and the angle formed by the position of the pointer with respect to the vertical and horizontal directions of the image showing the instrument panel can be obtained (5
8). After that, the image processing device compares the angular range of the normal value that the instrument may have set in advance for each instrument, and if the value indicated by the instrument is within that range, that value and "normal" If the signal is out of range, its value and an "abnormal" signal are sent to the alarm panel, which receives the signals and issues an alarm and records (59).

Next, with respect to the monitoring of the float type water level gauge showing the position of the water surface by the position of the float moving vertically, FIG.
This will be described in detail with reference to FIGS. 9 and 10. The image of the monitored device (floating water level gauge) captured by the visible camera is captured by the image processing device (61). This image is immediately copied (62) and noise-removed (63).

Thereafter, the determined initial threshold value (th)
Binarization processing (65) is performed by (64), and contraction / expansion processing (66) is performed to clarify the boundary line and to remove noise. Then, in order to find the position of the float (21 in FIG. 9) of the water gauge (22 in FIG. 9) and the proper height of the float in this image, labeling is performed on the object in the image (67). ), Among the many labeled objects, find the mark attached to the water gauge.

At this time, the area (a) and the perimeter (c) of each labeled object are measured (68). These values are compared with preset values (a ₀ , c ₀ ) (69), and if they are within the range, the following processing is performed, and if not, the threshold for binarization is set. The value is sequentially changed from the desired value, binarization processing is performed, and labeling is performed again (70).

After that, with respect to the patterns satisfying the above conditions, the angle θ formed by the principal axes of inertia of the screen with respect to the vertical and horizontal directions of the screen and the shape thereof are investigated (71,
72, 73) and specify a preset mark (74). In addition, the content of the flowchart described so far is a process to find the mark of the water level gauge in the captured image, so when capturing the water level gauge with a camera, it is necessary to select the camera position and angle for that purpose. is there.

The image which has been processed up to this point shows an image in which the mark attached to the water level gauge is white and the other marks are black. This mark indicates the range of appropriate values (y ₀ , y ₁ , x ₀ , x _{1 in} FIG. 10) of the position of the float (float) where the water level measured by the water level gauge is shown. Set the processing area as the reference.

First of all, the maximum and minimum positions (x ₀ , x ₁ ,
Search for y ₀ , y ₁ ) (75). Based on the coordinates obtained here, a window is set on the original screen of the image processed up to this point (the screen before binarization). The float to be detected appears dark on the screen, and generally the upper part is larger and the lower part is smaller.Therefore, the window area specified by the mark is divided into the same small areas by the dividing line perpendicular to the vertical direction in which the float moves. The average luminance is sequentially measured from the upper part of the screen for the small area (76), and the portion having the largest difference is detected as the upper surface position of the float (77).

This position is compared with the upper and lower limit positions of the appropriate float value obtained from the mark position information, and the height of the water level corresponding thereto is output. After that, the image processing device compares it with the normal value height range that would have been set for each water level gauge in advance, and if the value shown by the meter is within that range, it is compared with that value. If the signal is out of range, its value and an "abnormal" signal are sent to the alarm board, which receives the signals and provides an alarm and recording (78).

[0037]

[Effects of the Invention] By using or applying the monitoring apparatus and the monitoring method of the present invention to a conventional, for example, an abnormality monitoring robot, an abnormality monitoring robot can replace the conventional work of checking field instruments by a patrol officer. It is possible to prevent accidents. Further, also in the accident analysis after the occurrence of the accident, the detected values at the time of the accident of the various measuring instruments obtained by the device of the present invention can be effectively utilized.

[Brief description of drawings]

FIG. 1 is an overall configuration diagram of an embodiment of the present invention.

[Fig. 2]

[Fig. 3]

FIG. 4 is an explanatory diagram of a conventional technique.

[Fig. 5]

[FIG. 6]

FIG. 7 is an explanatory diagram of an embodiment in which the present invention is applied to read image processing of a circular instrument panel.

FIG. 8

FIG. 9

FIG. 10 is an explanatory diagram of an embodiment in which the present invention is applied to image processing for reading the height of a float of a water level gauge.

FIG. 11 is an explanatory diagram of contraction processing.

FIG. 12 is an explanatory diagram of expansion processing.

─────────────────────────────────────────────────── ─── Continuation of the front page (51) Int.Cl. ⁶ Identification code Internal reference number FI Technical display location G06T 7/20 G08B 25/00 510 M 7323-5G G08C 19/36 6964-2F

Claims (4)

[Claims] 1. A photographing device for photographing a monitored object and inputting an image thereof, an image processing judging device for judging whether a main body of the device equipped with the monitored object is normal or abnormal based on the inputted image, and the device. In a monitoring device provided with a device for displaying the judgment result of 1., a photographing means for photographing an area including a display surface of a measuring instrument provided in a plant device and inputting the image, and a binarizing process for the inputted image. Means for extracting a meter display surface image based on the feature amount of the meter display surface shape from the image after the above processing, information of the extracted meter display surface image and the original before binarization processing. Means for detecting the position of the pointer on the image of the display of the meter in the original image based on the image, and means for comparing the detected pointer position with the normal pointer position to determine if there is an abnormality and display it And a monitoring device provided with. 2. A step of photographing a monitored object and inputting an image thereof, a step of judging whether or not information indicated by the monitored object is normal based on the input image, and a step of displaying the judgment result. In the monitoring method having, a step of photographing an area including a display surface of a weighing instrument attached to a plant apparatus and inputting the image, and binarizing the input image and performing binarization from the binarized image. A step of removing noise, labeling each shape object included in the image after the above processing, and determining whether each labeling shape object corresponds to the display surface shape of the target weighing meter ,
A step of extracting a prometer meter display surface image from the image after the binarization processing, and determining an area of the prometer meter display surface from the image before the binarization processing based on the information of the extracted prometer meter display surface image. A step, a step of obtaining an average luminance for each of the small areas in the determined area, a step of detecting the position of the indicating means of the measuring instrument based on the average luminance, and a detected position of the indicating means. And a step of determining whether or not there is an abnormality based on the designated position at the time of normality and displaying the abnormality, and a monitoring method. 3. A monitoring device provided with a photographing device for photographing a monitored object and inputting an image thereof, and an image processing judging device for judging whether or not information indicated by the monitored object is normal based on the inputted image. In, means for inputting an image of an area including a display surface of a weighing meter provided in the plant device and inputting an image of the area, based on the geometrical characteristic amount of the display surface of the weighing meter A means for obtaining the display surface area, a means for obtaining the luminance distribution of the image within the obtained area, a means for obtaining the position of the indicating means of the measuring instrument based on the luminance distribution, and a means for showing the measuring instrument based on the position of the indicating means. A monitoring device provided with means for judging whether or not the information is within a normal range. 4. A step of photographing a monitored object and inputting an image thereof, a step of judging whether or not information indicated by the monitored object is normal based on the input image, and a step of outputting the judgment result. In the monitoring method including, a step of photographing an area including a display surface of a weighing instrument attached to a plant apparatus and inputting the image, a step of binarizing the input image, and a binarizing process. A step of determining a search range for the weighing meter display surface in the input original image based on the image, and a step of dividing the search range into a plurality of small areas to obtain an average luminance for each small area, The step of obtaining the position of the indicating means of the measuring instrument based on the obtained average luminance, the step of judging whether or not the information indicated by the measuring meter is normal based on the position information of the indicating means, and the step of outputting the judgment result. Specially equipped with Monitoring method to collect.