JPH02205998A - State monitoring device - Google Patents

Abstract

PURPOSE:To exactly detect only abnormality even when a processing objective picture is oscillated by executing picture processing to the whole processing objective picture, for which the state of an equipment is caught as a picture, further, manually setting the area of a processing picture, limiting the area of the picture to be the processing object, executing the picture processing and coupling these processed results. CONSTITUTION:The picture is displayed on a monitor, etc., and a window to be the processing objective area is manually set by a mouse, etc. Then, the window is controlled by a window control part 4 and the optimum window is set. After that, the picture processing is executed to the whole processing objective picture by a picture processing means 6 for the whole picture by using a parameter which is set by the window and a parameter control part. Next, the set window area is limited by a picture processing means 7 in the window and the picture processing is executed. Finally, these processed results are coupled by a picture processing coupling means and the presence and absence of the abnormality is detected from the result by an abnormality detecting means. Thus, the exact detection of the abnormality can be executed.

Description

[Detailed description of the invention]

(Industrial Application Field) The present invention relates to a condition monitoring device that monitors the condition of piping equipment, such as a thermal power plant, using an image processing device. [Prior Art] In a conventional condition monitoring device, For example, when monitoring piping equipment at a thermal power plant, etc., it is necessary to maintain the safety of the equipment by monitoring for leaks, damage, etc. Generally, such monitoring is carried out by patrolling or monitoring by patrol officers or operators. On the other hand, calculations are performed by comparing the current image with an image from a certain period of time ago against the overall image of the equipment status. A processing device performs calculations, and an abnormality is detected based on an abnormality occurrence signal from the processing device. [Problem to be solved by the invention] However, during plant operation, patrollers are present over a wide range of indoor and outdoor facilities. It is not easy for operators or operators to constantly monitor monitors showing the status of equipment; there are limits to the scope and content of monitoring, and oversights may occur due to oversights. Therefore, if an arithmetic processing unit calculates the overall image of the condition of the equipment and detects abnormalities, there is no need for operators to constantly monitor the equipment. When the image that captures the condition is shaken, there is a problem in that it is not possible to distinguish whether the abnormality occurring in a narrow range is vibration noise or the original abnormality. It is possible to reliably detect abnormalities occurring in a wide range to a narrow range that occur in the monitored image, even when the image is vibrating, and to continuously monitor equipment without the need for constant monitoring by patrol personnel or operators. The object of the present invention is to provide a condition monitoring device that can issue an appropriate alarm when an abnormality occurs during monitoring, and can display the condition of the equipment on a display device such as a monitor. ] The present invention captures the state of equipment as an image, displays the image or an image at a certain moment on a monitor, etc., and manually sets a window to be the processing target area using a mouse while viewing the image. The window management unit manages the optimal window, uses the window and the parameters set by the parameter management unit to perform image processing on the entire screen to be processed by the full-screen image processing means, and then Image processing is performed by limiting the set window area by the image processing means within the window, the results of these processes are combined by the image processing combination means, and the presence or absence of an abnormality is detected from this result by the abnormality detection means. If so, a warning is issued and an image is displayed.

[Effect]

According to the present invention, by manually setting a window to be processed while viewing an image displayed on a monitor, etc., the entire image that captures the state of the equipment is processed, and then further processing is performed. Since image processing is performed on a limited area and abnormalities are detected by combining these images, it is possible to accurately detect abnormalities that occur in a wide range, such as steam leaks, and abnormalities that occur in a narrow area, such as water/oil leaks. At this time, even if the image showing the state of the equipment is shaken, vibration noise is removed and only the abnormality is detected, so the state of the equipment can be reliably monitored. In addition, if an abnormality occurs, it is possible to issue an alarm and display an image.
It is possible to easily recognize where and what type of abnormality has occurred, greatly improving equipment maintenance. [Example] An example of the present invention will be described below with reference to the drawings. In the present invention, an arithmetic processing device may perform calculations on an image that captures the state of the equipment by combining one or more regions to be processed manually, or perform processing on the image. The calculation may be performed by a processing device by combining an image limited to a taught area with an image that does not specifically limit the target area and one measuring tool manually. A case will be described in which calculations are performed by combining images limited to a set area. FIG. 1 is a schematic configuration diagram of an embodiment device. In the figure 1
is an input section that displays the status of the equipment. An input unit consisting of an industrial television camera or the like captures an image of a monitoring target such as piping equipment and its surroundings, and displays this image on a display device 3 such as a monitor via an image data display unit 2. While viewing the monitoring target image displayed on the display device 3, the operator uses a mouse or the like to specify an area (window) for limiting the processing target using the window management unit 4. At this time, it is not necessary to specify any window at all, or one or more windows may be defined. The window management unit 4 also manages when an operator edits, such as changing, deleting, or adding to a defined window. On the other hand, various parameters are required to process one image. Regarding this parameter, in order to flexibly correspond to various images, the operator can specify the optimum value for the system using the parameter management section 5. The full screen image processing means 6 performs image processing on the entire screen of the image to be processed using window information from the window management section, parameter information from the parameter management section 7, and the like. Furthermore, the image processing means 7 in the window performs image processing by limiting the processing to the window set by the window management section 4, and combines the image processing means 6 for the entire screen and the image processing means 7 in the window into an image processing combination means. 8. Based on this result, the abnormality detection means 9 detects whether or not there is an abnormality, and if there is an abnormality, it is displayed on the display device 10. At this time, the image is also displayed on the display device 10. When performing status monitoring here, it is not necessary to create windows and parameters for images to be monitored each time the status monitoring system is started. They can be set manually by a designer or other person when constructing the system, and the worker can usually use the values set at that time as they are.If the values set at the time of system construction are not appropriate, the worker can change them. can do. FIG. 2 is a processing procedure diagram when a system designer sets windows and parameters during system construction and checks whether these set values are appropriate. Step 2
Step 1 determines which processing mode to perform, such as registration or modification, whether the window and parameters are appropriate, etc., and the input processing mode is determined in step 22. If the processing mode is registration, that is, if a new window, parameters, etc. are to be defined, an image name is input in step 23. If there are multiple objects to be monitored, any name that can be used to identify them may be used as the image name. If there is a single monitoring target, the image name may or may not be input. Step 23
In order to define a window corresponding to the image input in step 24 and perform image calculations in step 25, for example, 2.
Define parameters such as valorization thresholds. If the processing mode input in step 21 is modification, that is, if you want to modify or change already defined windows or parameters, etc., a list of existing image names is displayed in step 26, and from this list, in step 27 you can modify or modify. Specify the name of the image you want to change. Once the name of the image to be modified is determined, it is determined whether the window of the image is to be modified or the parameters are modified in step 28, and if the window is to be modified, the window of the image is changed in step 29. When modifying parameters, the parameters of the image are changed in step 30. Steps 24-25, Step 29-
In step 31, a check is made to see if the window and parameters entered and modified in step 30 are appropriate for the image to be monitored.
If it is not appropriate, the process is continued in step 32, steps 29-30 are repeated, and reconfirmation is made in step 31. In step 32, it is determined whether or not to continue the process, and if the process is to be continued, step 22
Return to If the window, parameters, etc. are appropriate and the setting is to be completed, it is decided in step 33 whether or not to store it in a file, and if it is to be stored in a file, storage processing is performed in step 34. When the end mode is specified in step 21, the process proceeds to step 33. FIG. 3 is a procedure diagram for managing window registration, modification, etc., and FIG. 4 is a diagram showing a method for specifying a window position. When registering or modifying a window, any device that can specify the coordinates of the area is fine, but
FIG. 3 is an example using a mouse with two buttons. In step 41, the left and right mouse buttons pressed are determined, and if the right button is pressed, the process ends. When the left button is pressed, step 42 determines whether a registered window exists within a certain distance from the coordinates of the position 51, and if it does not exist, the lower right coordinates 52 of the window are moved with the mouse in step 43. When specified, a rectangle 53 is drawn with the coordinates specified in step 41 as the top left coordinates of the window, and the window is registered in step 44. If it is determined in step 42 that a window exists within a certain distance, step In step 45, the operator inputs a processing command such as deletion or modification using the window as a processing target window, and in step 46, the window is modified or deleted. The method of determining the distance in step 42 is to
This is done by searching for the closest coordinates 55. If the window settings are to be continued, a determination is made in time for the operator in step 47, and if the process is to be continued, the process returns to step 41. Parameters to be input and modified in the parameter management unit 5 include the threshold for binarizing the difference image, the cumulative number of binary images, the number of shrinkage processes for the cumulative image, and the area for abnormality detection. Enter from. If these are not entered, the system will set standard values such as a binarization threshold value of 3, cumulative number of 10, number of contractions of 4, and area of 1. FIG. 5 is a table that stores values input and modified by the window management section 4. The number of windows 61 is the number of set windows in image name units, and indicates that this number of window information exists in the table. Table 62 stores the coordinates of window 1, and
63・yI41f! There is A64 and the lower right X coordinate 65 and X coordinate 66. Similarly, the coordinates of window 2 are stored in table 67 and the coordinates of window n are stored in table 68. n corresponds to the number of windows 61. FIG. 6 is a table that stores values input and modified by the parameter management section 5. Table 71 contains the threshold values for subtracting the basic image and the input image and binarizing the difference image, table 72 contains the cumulative number of binary images, and table 73
The table 74 stores the number of shrinkage processes to be performed on the accumulated images, and the table 74 stores the area for detecting an abnormality. If parameters other than these are required for image processing, they may be stored in the table successively. The information set in the window management section 4 and the parameter management section 5 does not particularly need to identify the processing object when image processing is performed on a single heavily taxed object. However, when a plurality of monitoring targets are photographed and double-view processing is performed, windows, parameters, etc. are often different for each monitoring target, so image processing data must be managed in accordance with the monitoring targets. Therefore, one person manually inputs the image name for each monitoring target using a keyboard or the like to identify it. 7th
The figure is a table diagram that connects image names and image processing data such as windows and parameters corresponding to the image names. The number of images 81 is the number of image names for identifying the monitoring target, and if the number of images per image is n, there are a table 82 indicating image 1, a table 85 indicating image 2, and a table 86 indicating image n below. The table 82 has a pointer 84 indicating a record in which the name 83 of the image 1, window, parameters, etc. are stored. FIG. 8 is a file configuration diagram when data such as windows and parameters are stored in a file in step 34 in correspondence with image names. The file contains the image name,
Next, the window and parameter information are stored in order for each image name. First, the number of images 81 is stored, and then image names and data pointers corresponding to the number of images are successively stored. The data pointer 84 indicates the first record in which the corresponding image data is stored.6 In order to perform image processing, the data is stored in the order of window information and then parameter information. The window information stores the number of windows, 61, and then stores the upper left coordinates and lower right coordinates for the number of windows. FIG. 8 shows an example when the number of windows is five. Parameter information includes binarization threshold 71, cumulative number 7
2. The number of contraction processes 73 and the area for abnormality 74 are stored in this order. When performing status monitoring, settings for windows, parameters, etc. do not have to be performed each time the device is started up. Once the monitoring target is determined, these values are also determined. Therefore, window and parameter information is usually monitored using the information entered and set by the system designer, etc., and only when changes are necessary, this information is entered or set.
Fix it. Window and parameter information set by the system designer etc. are stored in an external storage device such as a floppy disk in the configuration shown in Figure 8. During normal status monitoring, they are read from the external storage device and image processing is performed. . FIG. 9 is a procedure diagram of image processing in which image processing is performed only on the entire screen of the image to be monitored and within a designated window, and these are combined to detect an abnormality. In step 101, window and parameter data are read from an external storage device, and in step 102, an image of the monitoring target taken with an industrial television camera or the like of the input section 1 is fetched and this image is used as a reference image. Therefore, since the reference image is not an image prepared in advance, it can follow changes in brightness of the object to be monitored, making it easy to detect only abnormal areas. In step 103, the monitoring status is captured as an input image, and the difference with the reference image captured in step 102 is calculated. In step 104, the difference image is captured from an external storage device, binarized using the obtained threshold, and accumulated. do. Step 10
The reference image and input image to be processed in steps 2, 103, etc. are grayscale images with gradations, and the monitored object may be fluctuating due to the influence of motor vibration, etc. The cumulative image read from the external storage device in step 105 The number is compared with the number of times the difference image has been accumulated, and if it is less than or equal to the specified cumulative number, the process returns to step 103 to capture more input images, differentiate and accumulate them, otherwise perform the process of step 106, step 1
In step 06, the differentially accumulated image is shrunk in order to remove noise appearing around the image due to the influence of vibrations and the like. The number of contractions is the number of contractions read from the external storage device. The area of the shrunken image is determined in step 107, and the area for abnormality read from the external storage device is used in step 108.
If the calculated area is less than this value, step 1
At step 09, it is determined that there is no abnormality, and at step 110, it is determined that there is an abnormality otherwise. From step 103 to step 110,
Process the entire screen. Used to detect abnormalities that occur over a relatively wide area, such as steam leaks. However, abnormalities may occur over a wide area or within a narrow area. In the latter case, if the shrinkage process is performed in step 106, it will be removed and may not be detected as an abnormality. Therefore, further image processing is performed by limiting the area. In step 111, it is determined whether a window is defined. If so, a window is set in step 112, and in step 113, the area of the accumulated difference image is determined within the window. Step 114
Determine whether the area obtained in step 11 is less than a certain value (use the same value as in step 108), and if it is less than this value, proceed to step 11.
5, it is determined that there is no abnormality, and step 116, otherwise it is determined that there is an abnormality. In step 117, it is determined whether the defined window still exists, and if so, step 1
Return to step 12 and process the window in the same way. When the processing for all the set windows is completed, in step 118, the presence or absence of an abnormality on the entire screen in steps 109 and 110 is checked, and the presence or absence of an abnormality in the area limited to the window in steps 115 and 116 is checked in combination, and then in step 119 It is determined whether there is an abnormality in the processing on the whole screen or in the window, and if there is an abnormality in at least one of them, it is determined that there is an abnormality in the monitoring target in step 120 and that there is no abnormality in step 121. It is determined in step 122 whether there are any more monitoring targets, and if so, the process returns to step 102 to continue the process. In addition to the method of specifying the upper left coordinate and the lower right coordinate using a mouse or the like, the window of the present invention may be defined by specifying a changing point of the window. In short, all that is required is a rectangle to be defined. [Effects of the Invention] According to the present invention, image processing is performed on the entire screen of the image to be processed, which captures the state of the equipment as an image, and furthermore, the area of the image to be processed is manually set. Since we perform image processing with a limited number of images and combine them,
Even if the image to be processed is shaking, it is possible to accurately detect only abnormalities, greatly improving the reliability of monitoring. In addition, the condition of the equipment is captured as an image and calculated by a processing unit.
Since the presence or absence of an abnormality is detected from this result, there is no need for constant monitoring by patrol equipment or operators, and it is possible to prevent oversight of detection due to a decline in work efficiency due to long-term monitoring. Since the condition of the equipment can be monitored at any time, the safety of the equipment can be significantly improved.

[Brief explanation of the drawing]

The figures show one embodiment of the present invention. Fig. 1 is a schematic configuration diagram of the embodiment device, Fig. 2 is a processing procedure diagram for setting windows and parameters and checking them, and Fig. 3 is a diagram of the window and parameters. Processing procedure diagram for registration/correction, etc., Figure 4 shows how to specify the window position, Figure 5 shows the window table, Figure 6 shows the parameter table, and Figure 7 shows the image name. FIG. 8 is a diagram showing the configuration of files stored in the external storage device, and FIG. 9 is a processing procedure diagram for processing the entire screen and images in the window to detect an abnormality. 4... Window management section, 5... Parameter management section, 6... Full-screen image processing means, 7... Image processing means within the window, 8... Image processing combination means, 9.
...Anomaly detection means, 51...Upper left coordinates specified as a window, 52...Lower right coordinates specified as a window, 53...Window, 62...Window 1
9 Coordinate table, 82... Image 1 name tape. Figure 1 Collection Figure 2 Figure Figure 4 Figure 6 Figure 7 Figure 8

Claims (1)

[Scope of Claims] 1. In a condition monitoring device that includes an imaging device that captures the state of equipment as an image and an arithmetic processing device that processes the images of this imaging device, a range to be processed for images from the imaging device. a window management unit for limiting the image processing, a means for performing image processing on the entire screen from the imaging device, an image processing means for performing processing only within the limited range of the window, and an image processing unit for performing image processing on the entire screen. A condition monitoring device characterized by comprising means for combining image processing limited only within a window, and means for detecting an abnormality from the image processing results. 2. The condition monitoring device according to claim 1, wherein the window management function manually specifies the position of the window with respect to the image from the imaging device. 3. The image processing function for the entire screen or within a window for the image from the imaging device is characterized in that the difference between the reference image and the input image is accumulated and binarized, and the area of the binary image is determined. 1. The condition monitoring device according to 1. 4. The condition monitoring device according to claim 1, wherein an image captured at the start of monitoring without being prepared in advance is used as the reference image.