JPS6232336A - Liquid leak detecting method by itv camera - Google Patents

Abstract

PURPOSE:To detect automatically a liquid leaked from a valve, etc., and to reduce a burden on a plant monitoring person, by providing an image processor on a monitor system, and deciding a pool by processing a variable density image which is caught by an ITV camera. CONSTITUTION:A general purpose image processor 8 is provided additionally on a monitor system consisting of an ITV camera 3, an illumination use power source 4 and a monitor television 5. In such state, smoothing of a variable density image which is caught by the ITV camera 3 or smoothing which has held the edge is executed, and subsequently, an edge detection is executed to the image which is obtained by smoothing, by using eight kinds of differential operators. Next, the edge detected image is brought to a binary-coding processing by a prescribed threshold value, and thereafter, a variable density value of a boundary is extracted by executing a picture element expansion processing and a differential processing, and a pool is decided by detecting the existence of a sudden variation point. In such a way, a pool, etc. which have leaked from a pipe, a valve, etc. can be detected automatically.

Description

[Detailed Description of the Invention] [Technical Field of the Invention] The present invention provides an atomic couplant and a chemical plan!・Remove oil, water, etc. leaked from pipes or equipment such as valves.
This invention relates to a method of automatic detection using a 'T'V camera.

[Technical background of the invention and its problems]

BACKGROUND OF THE INVENTION Conventionally, vi devices such as piping or valves in an atomic reactor or a chemical plant have been monitored by a monitoring system using a closed-circuit television system such as an ITV camera.

FIG. 9 shows an example of the monitoring system, in which 1 is a pipe, 2 is a valve, 3 is an ITV camera, 4 is a light source for illumination,
5 is a monitor television.

In a monitoring system with such a configuration, if oil or water leaks from the valve 2 and a puddle 6 occurs on the floor surface 7, that is the only information that can be obtained, and based on that information, it is possible to determine whether or not a person is in the puddle. He was making such a judgment. Therefore, in this method, the plant monitor must constantly monitor the monitor television 5, which increases the burden on the plant monitor.
This reduced the ease of monitoring and operability of plant operations.

The present invention was made in view of the above circumstances, and its purpose is to automatically detect liquid leakage from piping, valves, etc. using an ITV camera, thereby reducing the burden on plant supervisors.

[Summary of the invention]

In order to achieve the above-mentioned object, the present invention is a monitoring device that monitors piping or valves in a plant using an ITV camera, in which a gray-scale image captured by the ITV camera is smoothed or smoothed while preserving edges. A smoothing process, an edge detection process in which edges are detected using an 8-way differential operator on the image obtained in this smoothing process, and an experimentally determined threshold for the image obtained in this edge detection process. A binarization process in which a value is binarized, a process to expand the binarized image obtained in this binarization process by several pixels, and a difference process is performed between the images before and after the expansion. a difference processing step, and a pale image extraction step for extracting the boundary gray values obtained in the difference processing step;
This method is characterized by comprising a step of determining whether there is a puddle by detecting the presence or absence of an abrupt change point in the 'a gray value obtained in the step of extracting the i11 light image.

[Embodiments of the invention]

Hereinafter, the present invention will be explained with reference to FIGS. 1 to 8.

FIG. 1 is a diagram showing an embodiment of the monitoring system according to the present invention, in which 1 is a pipe, 2 is a valve, 3 is an ITV camera, 4 is an illumination light source, 5 is a monitor TV, 6 is a water puddle, and 7 is a floor surface. , 8 are general-purpose image processing devices. Also, Figure 2 shows the IT
An example of a video m (image) of a puddle 6 captured by the V camera 3 is shown, and the pixel unit is expressed by 11 light values.

The vicinity of the boundary (edge) of the water puddle 6 has surface tension and is rounded. When the border is illuminated from the illumination light source 4 next to the ITV camera, brightness and darkness will occur at the border. This is a characteristic unique to puddles, and the present invention extracts this property to determine whether it is a puddle.

The extraction method will be explained below.

The image information obtained by the ITV camera 3 includes a light source, I
Noise generated by TV cameras, transmission lines, etc. is superimposed. To remove this noise, first use a 3x3 or 5x5 (3x3 is used in the following explanation) differential operator to smooth the image (original image) captured by the ITV camera 3 or smooth the image while preserving the edges. I do. Smoothing uses the same loading factor for the 3x3 operator and the second
A sum-of-products operation is performed between the original images shown in the figure. Also, there are several algorithms for smoothing that preserves edges.
The purpose is to perform smoothing while preserving contrast change points, that is, edges.

Next, edge detection is performed on the image obtained by the above smoothing using eight types of differential operators. The eight types of differential operators have loading coefficients as shown in Figure 3.
Edge detection is performed by taking the maximum value among the eight images processed by each operator. The edge detection image obtained in this manner is a grayscale image, and represents a point of sudden change in density. Note that the water puddle 6 has brightness and darkness, and since changes occur in the positive direction and the negative direction, the absolute value is taken.

Next, the edge detection image described above is subjected to a binarization process using an experimentally determined threshold value to create a binarized image shown in FIG. Thereafter, this binarized image is expanded by several pixels, and a difference process is performed between the images before and after expansion. Then, the difference image obtained by the difference processing is expanded to extract the boundary of the region.

Figure 5 shows the procedure for extracting the boundaries of an area.
(a) is an original image, (b) is a binarized image of the original image, (C) is an image expanded by several pixels, and (d) is a difference image. Also (e)
indicates an image obtained by extracting the difference image area from the original image,
In the figure, the black areas indicate boundaries, and the black areas and diagonally shaded areas are the values of the boundaries. In other words, this extraction extracts the boundary part of the puddle, and the contrast between light and dark puddles.

FIG. 6 shows the shading values of the AA section of the original picture shown in FIG. 5(a). In addition, Figures 7 and 8 show the procedure for extracting the shading values of the boundary area.
Part a) shows the original image of the pixel expansion processing image area and its gradation value, and part (b) shows the original image of the difference image area and its gradation value. As shown in FIGS. 6 and 7, by taking the difference between the image shown in FIG. 7(a) and the image shown in FIG. 7(b), the grayscale values shown in FIG. 8 are extracted. becomes possible. That is, gradation values in eight directions are extracted for this area, and the presence or absence of a sudden change point (brightness and darkness, which is a characteristic of a puddle) in the gradation values is detected to determine that it is a puddle.

〔Effect of the invention〕

As described above, according to the present invention, ambiguous information contained in the original image can be removed from a grayscale image captured by an ITV camera, and the characteristics of a puddle can be effectively extracted. Therefore, it is possible to automatically detect puddles of water leaking from pipes, valves, etc., and there is no need for Brandt monitors to constantly monitor the TV monitor as in the past, reducing the burden on the monitors and improving the ability to monitor plant operations. and operability can be improved.

[Brief explanation of drawings]

1 to 8 are diagrams for explaining the present invention, and the first
Figure 2 is a schematic diagram showing an embodiment of the monitoring system according to the present invention, Figure 2 is a diagram showing an image of a puddle captured by an ITV camera, Figure 3 is a diagram showing eight types of differential operators for edge detection, Figure 4 is a diagram showing a binarized image of a puddle;
Figure 5 is an explanatory diagram showing the procedure for extracting the boundary part of a puddle, Figure 6 is a diagram showing the shading values of the original image at the A-A cross section in Figure 5(a), and Figures 7 and 8 are FIG. 9 is an explanatory diagram showing a procedure for extracting the a-value of a boundary portion, and is a diagram showing an example of a conventional monitoring system. 1... Piping, 2... Bulb, 3... ITV camera, 4... Light source for illumination, 5... Monitor TV, 6...
- Water puddle, 7... Floor surface, 8... Image processing device. Applicant's representative Patent attorney Takehiko Suzue Figure 2\ I / Figure 3 Figure 4 (a) (b) (C)
(d) (e) Figure 5 Figure 6 Figure 7 Figure 8 Figure 9

Claims (1)

[Claims] In a monitoring device that monitors piping or valves, etc. in a plant using an ITV camera, a smoothing step is performed to smooth the grayscale image captured by the ITV camera or smooth the edge while preserving the edges, and an image obtained in this smoothing step. an edge detection step in which edges are detected using a differential operator in eight directions; a binarization processing step in which the image obtained in the edge detection step is binarized using an experimentally determined threshold; A difference processing step in which the binarized image obtained in this binarization processing step is expanded by several pixels, and a difference processing is performed between the images before and after the expansion, and a difference image obtained in this difference processing step. an expansion processing step in which the boundary is extracted by expanding the boundary, a grayscale image extraction step in which the grayscale value of the boundary obtained in this dilation processing step is extracted, and a sharp change in the grayscale value obtained in this grayscale image extraction step A method for detecting liquid leakage using an ITV camera, comprising the step of determining whether there is a puddle by detecting the presence or absence of a change point.