JP2001022433A - Monitoring device and monitoring method - Google Patents

Abstract

PROBLEM TO BE SOLVED: To highly precisely discriminate the scale/indicator of a measuring instrument, to precisely calculate an indicating value and to appropriately monitor the state of a plant in a monitoring device for reading the indicating value of the instrument as the parameter of abnormality monitoring and monitoring the plant based on the indicating value. SOLUTION: This monitoring device is provided with a photographing means for photographing an area including the indicating plane of a measuring instrument 1 installed in an apparatus to be monitored in a plant, a means 4 for taking in a photographed picture and deciding the position of the instrument 1 by a line feature value extraction filter processing and a binarization processing, a means for detecting the decided position of the indicator of the instrument 1 by the maximum value of the line feature value extraction filter processing, a judging means for judging whether the apparatus to be monitored is normal or not based on the position of the indicator, and a console 5 including a display means for displaying the result of judgment. Since it is not affected by the illuminance change due to illumination and shadow, the detection precision of an indicator area is improved and the occurrence of erroneous warning is reduced.

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 monitoring device for measuring and displaying pressure, temperature, flow rate, concentration, pH value, etc. in a device constituting a thermal power plant or the like. The present invention relates to a monitoring device and a monitoring method for reading a display value as a parameter for abnormality monitoring and monitoring a plant based on the read display value.

[0002]

2. Description of the Related Art In plants such as thermal power plants,
Due to demands for longer life of equipment and labor saving of patrol monitoring,
There is an increasing demand to automate equipment monitoring. for that reason,
2. Description of the Related Art A method is widely used in which, for example, an important device such as a fuel supply device, a burner, and a main valve in a boiler is photographed using a monitoring television camera or the like, and is visually observed on a monitor television for intensive monitoring.

[0003] Further, with the development of image processing technology, an example in which an image of a surveillance television camera or the like is image-processed to determine the presence or absence of an abnormality is spreading from FA (Factory Automation) to other fields, and is increasing. It is in.

[0004]

However, the surveillance television camera has a limited surveillance area and cannot cover the entire plant. In practice, the patrol observer patrols each part of the plant six to seven times a day. Important equipment is inspected visually or audibly. This patrol inspection work requires skill to find abnormalities, and places a considerable burden on patrol observers. On the other hand, the power plant aims to save labor costs by centralizing monitoring and saving labor.

FIG. 1 is a diagram showing an example of a procedure for detecting a display value of an instrument by a second conventional technique, that is, an image processing method. The image 6 of the circular instrument taken from a photographing device such as a surveillance television camera is sent to an image processing device, and binarized based on a predetermined threshold th. The binarization process is a process in which a pixel having a luminance equal to or less than the threshold th is set to white, and a pixel having a luminance exceeding the threshold th is set to black, and a white region and a black region are separated. The image obtained by the binarization processing of this method is a black-and-white image in which the luminance of the captured image is inverted and binarized.

[0006] The binarized image is labeled and divided into a plurality of scale areas and a pointer area 8. The labeling process is a process in which only white pixels are scanned, continuous pixels are set to the same region, and divided into a plurality of regions.

Next, a direction 10 having a maximum length 9 is determined for the pointer area 8 in the monitoring area 7 and the direction 10 and the horizontal line 1 are determined.
The inclination 11 formed by 2 is defined as the inclination of the pointer, and the indicated value of the pointer is read.

FIG. 2 is a diagram showing an example of erroneously determining a display value in the prior art. In the prior art shown in FIG. 1, when the luminance of an image changes due to a change in illuminance due to illumination, a shadow 13, or the like, in a binarization process based on a certain threshold th, the scale and the pointer are changed. The existing monitoring area 7 cannot be separated into the appropriate scale area and the pointer area 8;
It cannot be extracted as a scale and a pointer. Further, in the case of image input from a camera mounted on a camera platform or the like, the monitoring area 7 cannot be automatically set with high accuracy by a simple driving means, so that an image shift occurs due to the stopping direction of the camera platform.
The scale and pointer cannot be extracted.

In the above-mentioned prior art, there is no consideration for a change in illuminance due to illumination, shadow, or the like. Therefore, there is a problem that the detection accuracy of the pointer area 8 is low, and false alarms increase. In addition, when video from a monitoring television camera mounted on a driving device such as a pan head must be used, setting the monitoring area 7 with high accuracy requires a high stopping accuracy of the driving device, resulting in a large cost. It was a factor of up.

It is an object of the present invention to provide a monitoring device and a monitoring method capable of accurately distinguishing a scale and an indicator hand of an instrument of a plant, accurately calculating an indicated value, and appropriately monitoring a state of the plant. .

[0011]

SUMMARY OF THE INVENTION In order to achieve the above object, the present invention provides a photographing means for photographing an area including a display surface of an instrument installed in a device to be monitored of a plant, and a photographing means for photographing the photographed image. Means for determining the position of the instrument by line feature extraction filter processing and binarization processing; means for detecting the determined position of the pointer of the instrument by the maximum value of the line feature extraction filter processing; The present invention proposes a monitoring device including a determination unit that determines whether a monitoring target device is normal or abnormal based on the above, and a display unit that displays a result of the determination.

[0012] The present invention also provides an instrument which is provided on a device to be monitored of a plant and has a mark for position detection arranged on a display surface, an imaging means for photographing an area including the display surface of the instrument, and a photographed image. Means for determining the position of the instrument by detecting the position of the mark by the line feature extraction filter processing and the binarization processing, and detecting the determined position of the pointer of the instrument by the maximum value of the line feature extraction filter processing The present invention proposes a monitoring apparatus including: a unit that performs a determination based on the position of a pointer;

[0013] The present invention further captures an image of a region including a display surface of an instrument installed on a device to be monitored of a plant, and sequentially uses a local maximum value filter and a local minimum value filter by a predetermined number of times. A filter process for extracting a line feature amount of the image is executed, the image subjected to the local filter process is binarized, the binarized image is subjected to a labeling process, a shape coefficient is calculated for the labeled region, and a Determine the position, calculate the center of gravity for the determined instrument area, calculate the instrument width (height), set the monitoring area based on the center of gravity and the instrument width (height), and center the coordinates of the center of gravity Then, rotate the image from which the line features were extracted by a small angle,
The line feature in the monitoring area is measured, the angle at which the feature is maximized is detected, and the position of the rotation angle of the image at which the line feature is maximized is detected as the position of the pointer, indicated by the pointer. It is proposed to determine whether the display value of the plant exceeds the normal range and to display the status of the plant as normal or abnormal based on the result of the determination.

According to the present invention, a local maximum value filter and a local minimum value are captured by capturing an image of a region including a display surface of an instrument which is installed on a monitoring target device of a plant and has a plurality of marks for position detection arranged on the display surface. A filter process for extracting a line feature amount of an image of an instrument by sequentially using a value filter by a predetermined number of times is executed, an image subjected to local filtering is binarized, a binarized image is subjected to labeling, and labeling is performed. Calculate the shape factor for the determined area, specify the positions of multiple marks, determine the position of the instrument, calculate the center of gravity for the determined area of the instrument, calculate the width (height) of the instrument,
Set the monitoring area based on the center of gravity and the width (height) of the instrument, rotate the image from which the line features have been extracted by small angles around the coordinates of the center of gravity, and measure the line features in the monitoring area. The angle at which the amount is maximum is detected, the position of the rotation angle of the image at which the line feature value is maximum is detected as the position of the pointer, and whether the display value of the plant indicated by the pointer exceeds the normal range. It proposes a monitoring method of judging whether or not the plant status is normal or abnormal based on the judgment result.

In the present invention, the image processing apparatus executes a filtering process for extracting the line feature amount of the image of the instrument by using the local area filter, and outputs an image having only the line feature amount. Without being affected by changes in illuminance
The detection accuracy of the pointer area is improved, and false alarms are reduced.

Further, even when images from a surveillance television camera mounted on a driving device such as a camera platform must be used, the monitoring area can be set with high accuracy, and the driving device does not necessarily require high stopping accuracy. , Can reduce costs.

Further, in the present invention, for example, a circular gauge of the target image equipment taken out by image processing and a position of the pointer in a quarter area of the circular gauge are attached to the surface of each gauge. Calculated based on the mark,
Since the value indicated by the pointer of each instrument can be calculated accurately,
It is possible to remotely confirm the indicated value of the instrument, which was conventionally checked by the patrol observer, with sufficient accuracy.

[0018]

BEST MODE FOR CARRYING OUT THE INVENTION

FIG. 3 shows a monitoring apparatus according to a first embodiment of the present invention.
It is a perspective view which shows the whole structure of. The image of the circular instrument 1 captured from the photographing device 3 such as a surveillance television camera is
It is sent to the image processing device 4. The image processing device 4 includes the instrument 1
An image including the scales of the instrument 1 and the indicating needles necessary for reading the indicated value is extracted. The console 5 calculates the indicated value of the meter 1 based on the relative position between the scale of the meter 1 and the pointer in the extracted image.

FIG. 4 is a flow chart showing the first half of an example of the processing procedure for reading the pointer of a circular instrument according to the present invention in the first embodiment of FIG. 3, and FIG. 5 is a flowchart showing the pointer of the circular instrument according to the present invention. 9 is a flowchart illustrating a latter half of an example of a processing procedure for reading the data.

Step 14: The image processing device 4 captures an image of the circular instrument 1 photographed by the photographing device 3 such as a monitoring television camera.

Step 15: The image processing device 4 starts measuring the instrument 1
The filter processing for extracting the line feature amount of the image is executed, and an image including only the line feature amount is output.

FIG. 6 is a diagram for explaining a filtering process for extracting a line feature amount. As a line feature extraction filter,
Using a local maximum filter and a local minimum filter,
These two filter processes are sequentially executed by the specified number of times,
The difference between the images before and after the filter processing is extracted. The value of the line feature value increases as the feature of the line increases. Note that the width of a line that can be extracted increases in proportion to the designated number of times of the filter processing.

This local filter calculates the luminance for each of the pixels 35 to 44 constituting the local area, and calculates the maximum luminance or the minimum luminance of the calculation result as the pixel 4 at the center position.
The processing for determining the luminance of No. 3 is sequentially executed for the entire screen. When calculating the brightness, Equation (1) is used for the local maximum filter, and equation (2) is used for the local minimum filter. Here, A to I are parameters indicating the luminance of each of the nine pixels.

More specifically, when the sample image of the local region shown in FIG. 6 is processed, the pixel 4
The luminance 50 of 0 is the representative luminance of the pixel 43 at the center position.
Is the representative luminance of the pixel 43 at the center position.

Step 16: The image processing device 4 converts the image subjected to the local filter processing into a predetermined threshold value t
The binarization processing is performed according to h.

Step 17: The image processing device 4 executes a labeling process to find the circular instrument 1 from the binarized image. FIG. 7 is a diagram illustrating an example of the instrument outline mark 46 to be detected. Image processing device 4
Is a region 45 including the instrument outline mark 46 to be detected.
Is subjected to labeling processing.

Step 18: The image processing device 4 calculates a shape factor. For the labeled area, the shape factor shown in equation (3) is calculated. The shape coefficient is a coefficient obtained based on the relationship between the area of the labeled region and the perimeter, and according to the shape of the labeled region, ch = 4π (area of the region) / (perimeter of the region) ² … (3) The value is as follows.

Step 19: The image processing device 4 determines the magnitude of the shape factor. In the case of this embodiment, since a circular instrument is to be detected, chD = 0.85 and chU = 1.
Set to 0. In other words, if the shape factor exceeds 0.85, it is determined to be a candidate area for a circular instrument.

Step 20: When the region which is a candidate for a circular instrument is equal to or larger than the threshold value AU of a predetermined area, the image processing device 4 determines that region as the instrument position.

Step 21: If the shape factor is out of the predetermined range in Step 19, or if the area of the instrument candidate is small in Step 20, whether or not the number of regions for calculating the shape coefficient has reached a predetermined value. Check if. If the number of regions has not reached the predetermined value, the process returns to step 18 and the calculation of the shape coefficients of the regions is repeated.

Step 22: If the number of areas exceeds the predetermined value in step 21, the setting of the threshold value th is changed and the processing is repeated from the binarization processing in step 16. The above process is repeated until the position of the meter 1 can be determined in this way.

FIG. 8 is a diagram showing an image at each processing stage of the circular instrument read according to the present invention in the first embodiment.

Step 23: arithmetic means 5 in console
Determines the position of the instrument 1.

Step 24: The console 5 calculates the center of gravity 24 for the area where the position of the instrument 1 is determined, and sets its coordinates to (x, y).

Step 25: The console 5 calculates the width (height) 25 of the gauge 1.

Step 26: The console 5 has this width
A half of (height) 25 is a search range 47 in the height direction to be monitored, and the pointer is detected within this range.

Step 27: The console 5 calculates the center of gravity 24, the width (height) of the instrument 1, and the search range 47 calculated so far.
The monitoring area (7) is set based on. The monitoring area 2
7 can be arbitrarily set on the input screen.

Step 28: The console 5 moves to the center of gravity 24
The image from which the line feature value is extracted in step 15 is rotated by a small angle around the coordinates (x, y) of.

Step 29: The console 5 measures a line feature in the monitoring area 27.

Step 30: The console 5 repeats the above processing a predetermined number of times (n times) while measuring whether or not the rotation has been completed n times, and measures the characteristic amount at each rotation position.

Step 31: When the number of times reaches a predetermined number, the console 5 detects an angle at which the feature amount becomes maximum.

Step 32: The console 5 detects the position of the rotation angle of the image at which the line feature value has the maximum value as the position of the pointer. The line feature value has a higher value as the number of line portions increases. On the surface of the instrument 1, there is no component having a line other than the scale and the indicator needle. Among them, the indicator needle is the longest and thickest line. High value.

Step 33: The console 5 determines whether or not the display value of the plant indicated by the pointer exceeds the normal range, and displays the status of the plant as normal or abnormal.

Although the invention has been described with reference to an instrument having a circular outer shape, the invention of the first embodiment is also effective for an instrument having a square outer shape.

According to the first embodiment, the image processing apparatus executes the filtering process for extracting the line feature amount of the image of the instrument by using the local area filter, and outputs the image having only the line feature amount. Unaffected by changes in illuminance due to light and shadows,
The detection accuracy of the pointer area 8 is improved, and false alarms are reduced.

Further, even when an image from a surveillance television camera mounted on a driving device such as a pan head must be used, the monitoring area 7 can be set with high accuracy, and the driving device does not necessarily require high stopping accuracy. Cost can be reduced.

[0047]

FIG. 9 shows a monitoring apparatus according to a second embodiment of the present invention.
FIG. 3 is a diagram showing an example of the structure of a display surface of an instrument to be read in FIG. In the second embodiment, the position of the meter 1 is searched by detecting the mark 49 provided on the surface of the meter 1 instead of detecting the outer shape of the meter 1 of the first embodiment.

FIG. 10 is a flowchart showing the first half of an example of the processing procedure for reading the pointer of a square instrument according to the present invention in the second embodiment of FIG. 9, and FIG. 11 is a flowchart showing the present invention in the second embodiment of FIG. 6 is a flowchart showing the second half of an example of a processing procedure for reading a pointer of a square instrument by the following.

In the processing procedure of the second embodiment, the steps from the image input 14 to the labeling processing 18 are the same as those of the meter of the first embodiment. In the case of the instrument 1 as shown in FIG. 9, since the mark 49 for searching for the position of the instrument 1 is a square,
The shape factor calculated by the equation (3) is 0.75. Therefore, in step 19, chD = 0.7, chU =
0.8.

The position of the mark 49 is the rotation 2 of the pointer search.
8 is located at 50, 51 on the vertical and horizontal extensions from the center of gravity 24 (x, y).
Further, the distance from the center of gravity 24 (x, y) is set to be the same as the monitoring area 27 for searching for the pointer. As a result, the center of the rotation 28 for searching for the pointer
It becomes possible to set an area 27 for searching for the pointer to be (x, y).

Thereafter, the pointer can be detected by the same procedure as in the first embodiment. In step 34 of FIG. 10, it is determined whether the number is the second time or not because the number of marks 49 is two.
It corresponds to being an individual. If there are three marks,
The procedure may be changed so as to determine whether or not it is the third time.

Although the present invention has been described with reference to an instrument with a square mark, the invention of the second embodiment is also effective for an instrument with a circular mark.

According to the second embodiment, for example, a circular gauge of the target image equipment taken out by image processing and the position of the pointer in a quarter area of the circular gauge are attached to the surface of each gauge. Since the value indicated by the pointer of each instrument can be accurately calculated by using the mark 49 as a reference, it is possible to remotely confirm the indicated value of the instrument, which was conventionally confirmed by the patrol monitor, with sufficient accuracy.

The roles of the image processing apparatus 4 and the console 5 according to the present invention are not limited to the first and second embodiments. That is, in the flowcharts of FIGS. 4 and 5 or the flowcharts of FIGS. 10 and 11,
The function up to which step is provided in the image processing apparatus 4 and the subsequent steps are provided in the console 5 are arbitrary.

[0055]

According to the present invention, it is possible to execute a filter process for extracting a line feature amount of an image of an instrument using a local area filter and output an image of only the line feature amount. It is not affected by the change in illuminance, the detection accuracy of the pointer area is increased, and false alarms are reduced.

Further, even when an image from a surveillance television camera mounted on a driving device such as a camera platform must be used, the monitoring area can be set with high accuracy, and the driving device does not necessarily require high stopping accuracy. , Can reduce costs.

Further, the position of the pointer in the target image equipment taken out by image processing, for example, a circular instrument, and a quarter area of the circular instrument is determined with reference to the mark attached to the surface of each instrument. Since the value indicated by the pointer of each instrument can be accurately calculated, the indication value of the instrument, which has been conventionally checked by the patrol monitor, can be remotely confirmed with sufficient accuracy.

Therefore, by using or applying the present invention to a conventional surveillance robot or a campus camera, it is possible to substitute the monitoring robot or the campus camera for the on-site instrument checking work performed by the patrol observer. This can reduce the burden on patrol guards and reduce labor costs by centralizing monitoring and saving labor.

Also, in an accident analysis after an accident occurs, the clear detection values of the various instruments obtained from the monitoring device of the present invention at the time of the accident can be effectively utilized.

[Brief description of the drawings]

FIG. 1 is a diagram showing an example of a procedure for detecting a display value of an instrument by a conventional image processing method.

FIG. 2 is a diagram showing an example in which a display value is erroneously determined in the related art.

FIG. 3 is a perspective view illustrating an overall configuration of a monitoring device according to a first embodiment of the present invention.

FIG. 4 is a flowchart showing a first half of an example of a processing procedure for reading a pointer of a circular instrument according to the present invention in the first embodiment of FIG. 3;

5 is a flowchart showing the latter half of an example of a processing procedure for reading a pointer of a circular instrument according to the present invention in Embodiment 1 of FIG. 3;

FIG. 6 is a diagram illustrating filter processing for extracting a line feature amount.

FIG. 7 is a diagram showing an example of an instrument outer shape mark to be detected in the first embodiment of FIG. 3;

FIG. 8 is a diagram showing an image at each processing stage of a circular instrument read according to the present invention in the first embodiment of FIG. 3;

FIG. 9 is a diagram illustrating an example of a structure of a display surface of a meter to be read in the monitoring device according to the second embodiment of the present invention.

FIG. 10 is a flowchart illustrating a first half of an example of a processing procedure for reading a pointer of a square instrument according to the present invention in the second embodiment of FIG. 9;

FIG. 11 is a flowchart showing the second half of an example of a processing procedure for reading a pointer of a square instrument according to the present invention in the second embodiment of FIG. 9;

[Explanation of symbols]

 Reference Signs List 1 instrument 2 mark 3 visible camera 4 image processing device 5 console 6 image 7 monitoring area 8 area where scales and pointers exist 9 maximum length 10 directions 11 inclination 12 horizontal line 13 shadow 24 center of gravity 25 instrument width (height) 27 monitoring Area 28 Image rotation 35 to 43 Small area 44 Pixel center 45 Labeling area 46 Instrument outline mark 47 Height of search setting area 49 Mark 50 Extension line 51 Extension line

Continued on the front page (51) Int.Cl. ⁷ Identification symbol FI Theme coat II (reference) G08B 25/00 510 H04N 5/225 C 5H223 H04N 5/225 7/18 D 5L096 7/18 W G06F 15/70 330G F Terms (reference) 2F065 AA03 AA39 CC00 FF04 JJ03 JJ26 QQ05 QQ21 QQ29 QQ31 QQ33 2F076 BA14 BD05 BD07 BD11 BD14 BD16 BE05 BE09 5C022 AA03 AB65 AC01 AC69 5C054 AA01 CA04 CC03 EA01 EB05 EF08 FC02 DD00 FF04 FF19 FF30 GG02 GG19 GG21 GG23 GG29 GG31 5H223 AA01 BB01 CC01 DD09 EE01 FF03 5L096 BA02 BA18 CA05 DA02 DA03 EA43 FA03 FA14 FA60 FA64 GA34

Claims (4)

[Claims] 1. A photographing means for photographing an area including a display surface of an instrument installed on a device to be monitored of a plant, and taking the photographed image to obtain a position of the instrument by a line feature extraction filter process and a binarization process. Means for determining the position of the indicator hand determined by the maximum value of the line feature quantity extraction filter processing; and determining whether the monitored device is normal or abnormal based on the position of the indicator hand. A monitoring device comprising: a determination unit that performs the determination; and a display unit that displays a result of the determination. 2. An instrument installed on a device to be monitored of a plant, wherein a mark for position detection is arranged on a display surface, photographing means for photographing an area including the display surface of the instrument, and a line for taking in the photographed image. Means for detecting the position of the mark by a feature value extraction filter process and a binarization process to determine the position of the meter; and detecting the determined position of the pointer of the meter by the maximum value of the line feature value extraction filter process. A monitoring unit that determines whether the device to be monitored is normal or abnormal based on the position of the pointer, and a display that displays the result of the determination. 3. An image obtained by capturing an image of a region including a display surface of an instrument installed on a device to be monitored of a plant, and sequentially using a local maximum value filter and a local minimum value filter by a predetermined number of times. Executing a filtering process for extracting a feature amount, performing a binarization process on the locally filtered image, performing a labeling process on the binarized image, calculating a shape coefficient for the labeled region, and determining a position of the instrument. Determined, calculate the center of gravity for the determined area of the meter, calculate the width (height) of the meter, set the monitoring area based on the center of gravity, the width (height) of the meter, The image from which the line feature is extracted is rotated by a small angle around the coordinates, the line feature in the monitoring area is measured, the angle at which the feature is maximized is detected, and the line feature becomes the maximum value Image rotation angle The position of the degree is detected as the position of the pointer, and it is determined whether or not the display value of the plant indicated by the pointer exceeds the normal range. Based on the determination result, the state of the plant is displayed as normal or abnormal. How to monitor. 4. A local maximum value filter and a local minimum value filter which capture an image of a region including a display surface of an instrument which is provided on a monitoring target device of a plant and has a plurality of marks for position detection arranged on the display surface, Is sequentially used a predetermined number of times to execute a filtering process for extracting the line feature amount of the image of the instrument, binarizing the image subjected to the local filtering, labeling the binarized image, and labeling the image. The shape factor is calculated for the determined area, the positions of the plurality of marks are specified, the position of the meter is determined, the center of gravity is calculated for the determined area of the meter, and the width (height) of the meter is calculated. A monitoring area is set based on the center of gravity and the width (height) of the instrument. The image obtained by extracting the line feature amount is rotated by a small angle around the coordinates of the center of gravity, and the line feature amount in the monitoring area is calculated. The angle at which the feature value is maximum is detected, the position of the rotation angle of the image at which the line feature value is maximum is detected as the position of the pointer, and the display value of the plant indicated by the pointer is normal. A monitoring method for determining whether or not the range is exceeded, and displaying the status of the plant as normal or abnormal based on the result of the determination.