JPH07333039A - Method for measuring liquid level position using image processing - Google Patents

Abstract

PURPOSE:To automatically measure a height of a liquid level in a container by obtaining a border position between a side wall and the liquid level based on a relational position on an image between an image of a slant line drawn in advance on an inner side wall face of the container above the liquid level and a reflected shadow on the liquid level or a refracted image under the liquid level. CONSTITUTION:A slant line 24 and a standard length mark 27 for converting a size of an image on an image coordinate into a size of an actual spatial coordinate system are drawn in advance on an inner side wall face 15 of a container, and an upper end or a lower end of the mark 27 is aligned with a specific scale of a reference scale 28 for measuring a liquid level. With an image of a side wall 15 which has been picked by a monitor camera 1 processed, an intersection between an image of the line 24 above a liquid level 23 and a reflected shadow 25 of the line 24 on the level 23 or a refracted image 26 of the line 24 in transparent liquid is estimated to be a border (a position of the liquid level) 13 between the side wall 15 and the liquid level 23, and a coordinate on the image of the estimated border 13 is converted into an actual spatial coordinate by a conversion coefficient obtained by the mark 27. Thus the position of the actual liquid level 23 in the container can be immediately sensed.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an apparatus for automatically measuring the height of a liquid surface in a container such as a tank by image processing, and more particularly to position recognition by an image of the liquid surface.

[0002]

2. Description of the Related Art The height of a liquid surface (liquid level) measures the distance from a reference horizontal surface to the surface of a liquid. Conventionally, as a method for measuring this liquid level, the Mechanical Engineering Handbook (revised 6
Edition; Japan Society of Mechanical Engineers 1977; 6th edition Measuring method Chapter 7
As described in (7.6 Measurement of liquid level), as the measurement of displacement length, what is called a hook gauge or a point gauge is used to directly measure the liquid level, and a float is placed on the liquid surface. There was one that floated and mechanically measured the displacement of the float.

Further, as the measurement of pressure, the deformation of the diaphragm attached to the bottom surface and the side surface of the tank due to pressure is mechanically and electrically detected, or a thin tube is inserted into the liquid to make a tip. There was a method in which compressed air was ejected from the tube to measure the back pressure of the thin tube.

[0004]

Since the above-mentioned conventional mechanical liquid level detecting device cannot avoid mechanical failure, a person periodically makes a patrol to confirm that the measuring instrument is normal. There was a need. Therefore, there is a strong demand to eliminate such patrols in places where the liquid surface is to be checked in dangerous places or in distant places, and image processing is performed using images from surveillance cameras instead of conventional mechanical detection. The demand for automatic measurement of the liquid surface has increased. However, in order to measure the position of the liquid surface with an image, various difficulties are involved due to the nature of the liquid. That is, in the case of a highly viscous liquid, the liquid adheres to the wall surface near the liquid surface to a position higher than the liquid surface due to the capillary phenomenon, and it is difficult to determine the position of the liquid surface from the camera image. Further, in the case of a highly transparent liquid, the boundary of the liquid surface on the wall surface is unclear, and it is difficult to detect the liquid surface with an image.

An object of the present invention is to solve the problem of measuring the liquid surface by such an image and to provide a method capable of measuring the liquid surface position by processing the image from the monitor camera.

[0006]

The feature of the present invention is to analyze the condition on the side wall and the condition of the liquid surface in the image near the liquid surface captured by the camera, and determine the boundary position between the side wall and the upper surface of the liquid. By estimating the position of the upper surface of the liquid on the image, the calculated liquid surface position on the image is converted from the image coordinate system to the actual spatial coordinate system to determine the actual position of the liquid surface. There is something I did.

Specifically, an oblique straight line is drawn in advance on the wall surface, and the positional relationship on the image between the straight line image of the liquid taken in by the camera and the reflected image on the liquid surface is shown. Based on this, the boundary position between the side wall and the upper surface of the liquid is determined.

Further, an oblique straight line is drawn in advance on the wall surface, and an image of the straight line image of the liquid taken in by the camera and the straight line refraction image in the liquid, that is, below the liquid surface, is formed on the image. The boundary position between the side wall and the upper surface of the liquid is determined based on the positional relationship of.

[0009]

According to the present invention, an oblique straight line previously drawn on the wall surface is used, and the coordinates of the intersection of the reflected shadow of the straight line and the straight line above the liquid surface are obtained on the image, and the liquid surface is based on this intersection. Therefore, if the liquid has a low transparency and the surface has some reflectance, it is possible to obtain the position of the liquid surface by image processing.

Furthermore, according to the present invention, an oblique straight line previously drawn on the wall surface is used, and the coordinates of the intersection of the straight line below the liquid surface and the straight line above the liquid surface are obtained on the image. Since the position of the liquid surface is obtained based on the intersection, the position of the liquid surface can be obtained by image processing even in the case of a highly transparent liquid.

[0011]

Embodiments of the present invention will be described below with reference to the drawings.

FIG. 1 schematically shows the overall structure of the liquid surface position measuring apparatus according to one embodiment and the principle of measurement. Here, as an example, a device for monitoring the liquid level 23 in the tank 11 is described, but for the purpose of monitoring the liquid level, in an artificial liquid container, the liquid level of a liquid existing in nature, etc. Shall apply in the same manner. The camera 1 captures an image of the side wall 15 portion of the tank 11 to be monitored, and the camera image is transmitted to the tank central liquid level monitoring station 12 via the optical cable 7. At the central liquid level monitoring station 12 in the tank, an image is received by the photoelectric converter 6, converted into an electric signal, and transferred to the video switch 8. The video switching device 8 distributes the image signal to the monitoring monitor 3 and the image processing device 2. The monitoring monitor 3 is a conventional display device for a human monitor, and the feature of the present invention resides in that an image processing device is added. The image processing device 2 automatically measures the position of the liquid surface by the image processing method peculiar to the present invention, and transfers the result to the data output management device 4. With such a configuration, the liquid level in the tank at the site can be measured without using any mechanical measuring instrument.
The image is automatically measured by the image processing device 2 and stored in the database 9 by the data output management device 4, or the printer 10
Can output a hard copy or send or provide it to the required department. The most characteristic feature of the present invention is a method of detecting the liquid surface position by an image, and the details thereof will be described below.

FIG. 2 shows an object 18 of the wall surface image pickup unit 15 and a reflection shadow 19 on the liquid surface near the image processing target when the reflectance of the surface of the liquid surface is good, through the lens 16 of the camera 1. It shows how it is captured on the imaging plate 17. When viewed from the camera side, the reflection shadow 19 on the liquid surface of the object 18 is created as shown in FIG. 2 due to the geometrical relationship with the lens 16 of the camera. The reflection shadow 19 looks equivalent to the image 20 formed on the opposite side of the object 18 when viewed from the camera side.
The object 18 and the image 20 have the same size and shape, and the present invention utilizes this feature. It should be noted that the image of the object 18 on the imaging plate is 21, the image of the image 20 on the imaging plate is 22, and the images 21 and 22 are images of the same size. When monitoring the liquid surface, the boundary between the liquid surface and the side wall 13
It is only necessary to be able to directly recognize, but it is generally difficult.
There are various reasons, but the biggest reason is that it is due to the capillary phenomenon and that the liquid is transparent. Due to the capillarity, the liquid rises along the side wall near the boundary 13, and the portion where the liquid has risen easily absorbs light and changes its color to black. For this reason, the entire boundary portion becomes black, and the boundary of the liquid surface cannot be recognized on the image at all. Even if such a phenomenon is unlikely to occur, if the liquid has strong transparency, the sidewall 1
The boundary 13 between the liquid surface 5 and the liquid surface 23 is, of course, almost unrecognizable on the image. FIG. 3 is an example in which a straight line (line segment) 24 is drawn on the wall surface so that the height of the surface of such a liquid can be recognized by image processing. In this example, when the liquid has high transparency, the straight line portion below the liquid surface can be captured as a refraction image 26 on the camera image, and since the transparency is high, a good image that can be image-processed can be obtained. The reflection shadow 25 of the line segment 24 seen above the liquid surface on the liquid surface may be caught on the image, but in this case the reflectance is poor and a good image that can be recognized by image processing cannot be obtained. Here is an example:

FIG. 4 shows an example in which, contrary to FIG. 3, the transparency is poor and the surface reflectance is good instead. In this case, the refraction image 26 below the liquid surface is not a good image, but the reflection shadow 25 on the liquid surface of the line segment 24 above the liquid surface.
Is an example in which a good image is obtained. 3 and 4 together,
The boundary portion 13 between the liquid surface 23 and the side wall 15 is an example that cannot be directly recognized in an image due to a capillary phenomenon or the like. In the case of FIG. 3, the image 26
In the case of FIG. 4, the image 25 can be successfully used to estimate the position of the boundary. FIG. 5 shows an embodiment in which an inclined straight line 24 is drawn on the wall surface 15 and the height position of the liquid surface is estimated using the refraction image 26 and the reflection shadow 25 on the liquid surface 23. 27
Is a mark (which may be attached on the scale) for converting the size of the image in the image coordinates into the size in the actual spatial coordinate system, which is drawn in advance on the wall surface. Here, it is assumed that a mark with a height of Y1 mm (long rectangular shape) is attached. The position of the lower end or the upper end of this mark is made to coincide with a specific memory on the reference scale 28 for measuring this liquid level (the lower end is made to coincide with the position of 8 cm in this figure). By arranging in this way, the comparison with the reference can be easily performed by image processing, so that the position of the liquid surface can be calculated in the actual spatial coordinate system as soon as the boundary position 13 is detected. A method of obtaining the conversion coefficient between the image coordinate system and the actual spatial coordinate system will be described with reference to FIG. First, an image including a standard length (Y1 mm) mark is captured (block A). Next, the image of the standard length mark portion is processed to measure its length y2 in the image coordinates (block B). Finally, the conversion coefficient K is calculated by the following formula (block C).

K = Y1 / y2 The conversion coefficient K does not have to be calculated all the time, but may be calculated and stored once, and it is recalculated when the conditions change or the stored one disappears. Therefore, there is no problem even if the mark is temporarily partially or entirely submerged in water.

Next, an embodiment in the case of a liquid having a good transparency will be described with reference to FIGS. 3, 7 and 9. In this case, as shown in FIG. 3, the refraction image 26 of the straight line below the liquid surface is clear, but the reflection shadow 25 on the liquid surface is not clear. Therefore, the straight line 24 and the refraction image 26 are image-processed to estimate the liquid surface height. The straight line 24 above the liquid and the refraction image 26 below the liquid show a state of being bent by the refractive index of the liquid as shown in FIG. Use this feature. A method of obtaining the liquid surface position 13 as an intersection of a straight line 24 on the liquid surface and a refraction straight line 26 below the liquid surface will be described with reference to FIG. First, an image is captured (block A), and a straight line (line segment) 24 above the liquid surface is obtained. The calculated straight line is defined by the following equation (block B).

Y = A1 * x + B1 The refraction line 26 below the liquid surface is determined. The obtained straight line is defined by the following equation (block C).

Y = A2 * x + B2 The intersection (xp, yp) of the two straight lines is obtained from the following equation (block D).

[0019]

[Equation 1]

Further, the position (xr, y) of the standard length mark
r) is obtained on the image (block E). The difference in the Y direction between the position (xr, yr) of the standard length mark and the intersection (xp, yp) is calculated (block F).

H '= yr-yp The actual distance H between the mark having the standard length and the intersection is obtained by the distance conversion coefficient K (block G).

H = K · H 'Since the height (Ho) of the standard length mark is known, the liquid level height is calculated as liquid level height = Ho-H (block H). Here, a method of obtaining the intersection of the straight line 24 and the straight line 26 by image processing will be described in detail with reference to FIG. In FIG. 3, the upper end of the straight line 24 is P3, and the lower end of the line segment 26 is P1. The intersection of two straight lines is P2
And The intersection P2 is easily obtained when two straight lines are obtained. When obtaining the line segment 24 by image processing, two windows W3 and W4 are provided on the line segment, the Y-axis projection cumulative histogram is calculated in this window, and the position of the intersection of the window and the straight line is obtained. An accurate straight line 24 can be obtained by connecting two points obtained by the histogram calculation in the two windows with a straight line. The same applies to the straight line 26. Two windows W1 and W2 are created on the straight line, the Y-axis projection cumulative histogram is calculated in the window, two intersections of the window and the straight line are obtained, and these two intersections are formed by the straight line. A straight line 26 is obtained by tying. Straight line 2
Since 6 and the intersection point P2 change due to the change in the liquid level of the liquid, it is necessary to know the outlines of the two line segments in advance in order to create four windows. The processing procedure will be described below with reference to FIG. First, the Y-axis projection cumulative histogram 34 is created for the entire input image (box A). In the above-mentioned histogram 34, since the change of the point P1 and the Y-axis projected portion of P3 is large due to the influence of the straight lines 24 and 26, P1,
The approximate y-coordinates P1y and P3y of the point P3 can be easily obtained (box B). Next, the X-axis projection cumulative histogram 35 is created for the entire input image (box C). In the histogram 35, the changes in the X-axis projection portion between the point P1 and P3 are large due to the influence of the straight line 24 and the straight line 26.
Three approximate x-coordinates P1x, P3x are obtained (box D). From the approximate values of the points P1 and P3, the approximate value of the point P2 is obtained as follows. The straight line 24 is the point P3 (P3X, P3
y) and the angle with respect to the Y axis is i1. Slope K
When expressed by 24, K24 = tan (90-i1). i1 is known.

On the other hand, the straight line 26 is a point P1 (P1X, P1
y) and is at an angle i2 with respect to the Y axis. Slope K2
When expressed by 6, K26 = tan (90-i2). i2 is determined from the refractive index n of the liquid and i1.

From Snell's law, the following relationship is established, where n is the refractive index of the liquid.

N = sin (i2) / sin (i1) Therefore, since i1 and n are known, i2 is also known. Therefore, the slopes of the two straight lines 24, 26 are known.

Therefore, the straight line 24 can be expressed by the following equation.

Y-P3y = K24 (x-P3x) or y-K24 * x = P3y-K24 * P3x The straight line 26 can be expressed by the following equation.

Y-P1y = K26 (x-P1x) or y-K26 * x = P1y-K26 * P1x The intersection of the two straight lines is the P2 point and is represented by the following equation.

[0029]

[Equation 2]

The coordinates (P2x, P2y) of the point P2 described above are merely approximate values and do not give an accurate position of the liquid surface. To measure the exact position of the liquid surface,
Using the approximate positions of 1, P2 and P3, the window W
1, W2, W3, W4 are created on a straight line, histogram processing is performed in each of the two sets of windows, the intersection of the window and the straight line is obtained, and the two straight lines are connected to form a display formula for each straight line. obtain. Windows W3 and W4 on line 24
The procedure for creating the will be described. Both ends P of line segment 24
Since the approximate coordinates of the points P2 and P3 are known, an appropriate internal division point of the line segment may be calculated and two windows may be created at the points. For example, if the line segments P2 and P3 are divided into four equal parts and the interior division points are named Pn1, Pn2 and Pn3 in order from the left,
It is preferable to place the window W3 at the Pn1 point and the window W4 at the Pn3 point.

The coordinates of the window in this case are as follows.

Coordinates of window W3 (W3x, W3y) W3x = (3 * P2x + P3x) / 4 W3y = (3 * P2y + P3y) / 4 Coordinates of window W4 (W4x, W4y) W4x = (P2x + 3 * P3x) / 4 W4y = (P2y + 3 * P3y) / 4 Similarly, it is necessary to create the windows W1 and W2 on the straight line 26 as well. The creation procedure is similar to that described above. The coordinates of the window in this case are as follows.

Coordinates of window W1 (W1x, W1y) W1x = (3 * P1x + P2x) / 4 W1y = (3 * P1y + P2y) / 4 Coordinates of window W2 (W2x, W2y) W2x = (P1x + 3 * P2x) / 4 W2y = (P1y + 3 * P2y) / 4 In this way, two windows each were created on the two straight lines. When a Y-axis projection cumulative histogram is created in each window, the position of the intersection between the window and the straight line can be obtained because the histogram changes greatly at the intersection with the straight line. Since the straight line is thick in this case, the positions of both the upper edge and the lower edge of the straight line can be obtained. In this case, only the upper edge information is used and the lower edge information is discarded. Set the Y coordinate of the intersection of the windows W1, W2, W3, W4 and the straight line to W
1y, W2y, W3y, W4y.

The straight line 24 is (W3x, W3y) and (W4
x, W4y) is obtained from the following equation as a straight line.

Y = (W4y-W3y) (x-W3x) /
(W4x−W3x) + W3y The straight line 26 has (W1x, W1y) and (W2x, W2y).
It is obtained as the straight line connecting

Y = (W1y-W2y) (x-W2x) /
(W1x−W2x) + W2y Next, from FIG. 4, FIG. 8 and FIG. 10, in the case of a liquid having poor transparency and relatively good surface reflectance, a straight line 24 on the liquid surface
An embodiment using the reflection shadow 25 of will be described. In this case, as shown in FIG. 4, a refraction image 26 of a straight line below the liquid surface is obtained.
Is not clear, but the reflection shadow 25 on the liquid surface has a feature of being clear. Therefore, the reflected image 25 is image-processed to estimate the liquid surface height. The straight line 24 and the reflected image 25 above the liquid are symmetrical figures with respect to the liquid surface position 13 as shown in FIG. Use this feature. A method of obtaining the liquid surface position 13 as an intersection of a straight line 24 on the liquid surface and a straight line 25 of the reflection image on the liquid surface, or calculating the Y coordinate on both straight lines at the position where the X coordinate is the same and the center thereof There is a method that uses the value as the liquid surface position. Either method may be selected in view of ease of processing, good accuracy, and the like. Here, the method of obtaining the intersection of two straight lines will be described in detail with reference to FIG. First, an image is captured (box A), and the upper straight line 24 is obtained (box B). Next, the straight line 25 of the reflection image on the lower side is obtained (box C). Find the intersection (xp, yp) of the two straight lines (box D). Further, the position (xr, yr) of the standard length mark is obtained on the image (box E). The difference in height between the intersection (xp, yp) and the standard length mark position (xr, yr) is calculated (box F), and the actual distance H is obtained by the distance conversion coefficient K (box G). Since the height (Ho) of the standard length mark is known, the liquid level height is calculated as Ho-H (box H).

Here, according to FIG. 10, line segment 24 and line segment 25
The method of finding the intersection point will be described in detail. In FIG. 4, the upper end of the line segment 24 is P3 'and the lower end of the line segment 25 is P4'. The intersection of the two line segments is P2 '. The intersection P2 'can be easily obtained when two straight lines are obtained. When the line segment 24 is obtained by image processing, two windows W3 'and W4' are provided on the line segment and the Y-axis projection cumulative histogram is calculated in this window. An accurate straight line 24 can be obtained by connecting the two points obtained by the histogram calculation in the two windows. The same applies to the line segment 25, and two windows W1 'and W2' are created on the line segment 25 and obtained in the same manner. Since the intersection point P2 'changes due to the change in the liquid level of the liquid, it is necessary to know the outline of the two straight lines in advance in order to create four windows. The processing procedure will be described below with reference to FIG.
First, the Y-axis projection cumulative histogram 36 is created for the entire input image (box A). Histogram 3 above
6 is the influence of the line segment 24 and the line segment 25, and the points P3 'and P4'
Since there is a large change in the Y-axis projection portion of, the approximate y-coordinates P3y 'and P4y' of the points P3 'and P4' are obtained (box B). Next, the X-axis projection cumulative histogram 37 is created for the entire input image (box C). The histogram 37 has P2 'and P due to the influence of the line segment 24 and the line segment 25.
Since the change of the X-axis projection portion of 3'is large, approximate x-coordinates P2x 'and P3x' of the points P2 'and P3' are obtained (box D). P4x 'is equal to P3x'. The Y coordinate of the point P2 'is obtained as an intermediate value of the Y coordinates of the points P3' and P4 '.

P2y '= (P3y' + P4y ') / 2 As described above, P2' (P2x ', P2y'), P3 '(P
3x ', P3y'), P4 '(P4x', P4y ') were all obtained.

In order to obtain the line segments 24 and 25 more accurately, two windows W3 'and W4' are created in the line segment 24. Since both ends P2 'and P3' of the line segment 24 are known, a window can be easily created on the line segment 24 using this. If the above-mentioned window is created by dividing the line segment 24 into four equal parts, the coordinates of the window in this case are as follows.

The coordinates of the window W3 '(W3x', W3
y ′) W3x ′ = (3 * P2x ′ + P3x ′) / 4 W3y ′ = (3 * P2y ′ + P3y ′) / 4 Coordinates (W4x ′, W4y ′) of the window W4 ′ W4x ′ = (P2x ′ + 3 * P3x) ′) / 4 W4y ′ = (P2y ′ + 3 * P3y ′) / 4 Similarly, it is necessary to create windows W1 ′ and W2 ′ on the line segment 25 as well. The creation procedure is similar to that described above. The coordinates of the window in this case are as follows.

The coordinates of the window W1 '(W1x', W1
y ′) W1x ′ = (3 * P2x ′ + P4x ′) / 4 W1y ′ = (3 * P2y ′ + P4y ′) / 4 Coordinates (W2x ′, W2y ′) of the window W2 ′ W2x ′ = (P2x ′ + 3 * P4x) ′) / 4 W2y ′ = (P2y ′ + 3 * P4y ′) / 4 In this way, two windows each were created on the two line segments. When a Y-axis projection cumulative histogram is created in each window, the position on the line segment can be obtained because the change in the histogram is large at the portion intersecting the line segment. Since the line segment has a thickness in this case, the positions of the upper edge and the lower edge of the line segment can be obtained. In the case of the windows W3 'and W4', it is preferable to use only the information of the upper edge portion, and in the case of the windows W1 'and W2', it is preferable to use only the information of the lower edge portion. Y at the intersection of the window W1 ', W2', W3 ', W4' and the line segment
The coordinates are W1y ', W2y', W3y ', W4y'.

The line segment 24 is obtained from the following equation as a straight line connecting (W2x ', W2y') and (W3x ', W3y').

Y = (W3y'-W2y ') (x-W2x')
/ (W3x'-W2x ') + W2y' The line segment 26 has (W1x ', W1y') and (W2x ', W
It is obtained from the following equation as a straight line connecting 2y ').

Y = (W1y'-W2y ') (x-W2x')
/ (W1x'-W2x ') + W2y' The liquid level height can be obtained as the intersection of the above two straight lines.
Described earlier.

As described above, according to the present invention, a straight line inclined to the side wall is added in order to overcome the unclear point of the liquid surface due to the capillary phenomenon and the transparency which are the peculiar properties of the liquid. It became possible to automatically measure the surface height of the liquid.

FIG. 11 shows an example of measuring the liquid level when the transparency of the liquid is not good. There is a scale (memory and numbers) on the wall, but this is the method to read the scale to the point where it can be read by image processing and use that point as the liquid level.

FIG. 12 shows an embodiment of the invention in the case where the side wall is discolored by the capillary phenomenon and the boundary portion of the liquid surface cannot be recognized as an image. Point A is the boundary 13 between the liquid surface 23 and the side wall 15.
In FIG. 13, the liquid level measurement is performed in the case where the color changes from point A to point A ′ due to the capillary phenomenon and the AA ′ portion is reflected on the liquid surface and has a reflected image AA ″. An example of the image processing of will be described below.
The positions of the points A ′ and A ″ are measured in the image. The midpoint between the points A ′ and A ″ is calculated to obtain the position of the point A and the liquid surface.

FIG. 14 shows an embodiment in which the projection 33 is present in the liquid. The liquid is required to have a low transparency and a high surface reflectance. Liquid upper part 33a of protrusion
The part has an image such as 33b reflected in the surface of the liquid in the image, and the image of the part 33c in the liquid cannot be seen well because the transparency is not good. An example of image processing in such a case is shown in FIG.
Shown in. First, the image is captured and the height y of the portions 33a and 33b is measured in the image. If y is converted to an actual dimension and Y = K * y and the position H0 of the top of the protrusion is known, the position of the liquid surface is
It is obtained by the following formula.

H = H0-Y / 2

[0050]

According to the present invention, even when the boundary between the wall surface and the liquid surface is unclear due to the capillary phenomenon or the transparent nature of the liquid, the height of the liquid surface can be determined by the image processing, and the handling is easy. It becomes possible to automatically measure the liquid level by easy image processing. As a result, automatic measurement can be performed simply by installing a surveillance camera on the site, and periodic patrol of people is not required.

[Brief description of drawings]

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

FIG. 2 is a diagram illustrating the principle of an image appearing below the liquid surface due to the reflection of the upper surface of the liquid surface.

FIG. 3 is an explanatory diagram of a method using an inclined straight line (when the transparency of the liquid is good).

FIG. 4 is an explanatory diagram of a system using an inclined straight line (when the liquid has poor transparency and good surface reflectance).

FIG. 5 is an explanatory diagram of a method that uses a reflection shadow of a surface of an inclined straight line.

FIG. 6 is a flow chart for obtaining a conversion coefficient between the actual length and the length in the image.

FIG. 7 is a flow chart for measuring a liquid level height using a linear refraction image below the liquid level.

FIG. 8 is a flow chart for measuring the height of a liquid surface using a reflection shadow of a straight line on the liquid surface.

FIG. 9 is a processing flow chart for obtaining a straight line and its intersection.

FIG. 10 is a processing flow chart for obtaining a straight line and an intersection thereof.

FIG. 11 is a flow chart for reading the scale and characters and measuring the height of the liquid surface.

FIG. 12 is a flow chart for measuring a liquid surface height using a reflection shadow of a portion that is wet and discolored by a capillary phenomenon between a liquid surface and a side wall boundary portion.

FIG. 13 is a flow chart for measuring a liquid level height using a liquid wetting discolored surface.

FIG. 14 is an explanatory diagram of another example in which the liquid surface is a mirror surface and projections are present in the liquid.

FIG. 15 is a flow chart of image processing when the liquid surface is a mirror surface and projections are present in the liquid.

[Explanation of symbols]

1 ... Camera, 2 ... Image processing device, 3 ... Monitoring monitor T
V, 4 ... Data output management device, 5 ... Electro-optical converter, 6 ... Photoelectric converter, 7 ... Optical cable, 8 ... Image switching device, 9 ... Database, 10 ... Printer, 11 ... Tank, 12 ... Tank liquid level center Monitoring station, 13 ... Boundary between wall surface and liquid surface, 15 ...
Side wall.

 ─────────────────────────────────────────────────── ─── Continuation of the front page (72) Inventor Masao Takafuji 7-1-1 Omika-cho, Hitachi-shi, Ibaraki Hitachi Research Laboratory, Hitachi Ltd. (72) Inventor Hirotomo Hotta 2-chome, Misaki-cho, Chiyoda-ku, Tokyo 9-18 No. 18 within Hitachi System Technology Co., Ltd.

Claims (7)

[Claims] 1. An ITV camera for capturing an image of a surface of a liquid to be measured, and an image processing device for capturing an image from the camera and performing image analysis,
Image analysis of the situation on the side wall and the state of the liquid surface in the image captured from the camera, to determine the position of the upper surface of the liquid on the image by estimating the boundary position of the side wall and the liquid upper surface, A liquid surface position measuring method using image processing, characterized in that the liquid surface position on the obtained image is converted from an image coordinate system to an actual spatial coordinate system to obtain an actual position of the liquid surface. 2. The method according to claim 1, wherein an oblique straight line is drawn on the side wall in advance, or an object is attached to obtain an oblique straight line in the image, and the image of the straight line above the liquid and the A liquid surface position measuring method using image processing, characterized in that a boundary position between the side wall and the liquid upper surface is estimated based on a positional relationship between the image and the image. 3. An image of a straight line above the liquid according to claim 1, wherein a diagonal straight line is drawn in advance on the side wall or an object is attached to obtain a diagonal straight line,
A liquid surface position measuring method using image processing, characterized in that a boundary position between the side wall and the liquid upper surface is estimated based on a positional relationship on the image with a straight reflection image on the liquid surface above the liquid. 4. The liquid surface of claim 1, wherein a mark having a standard length is attached to the side wall in advance, and the length of the mark on the image is converted from an image coordinate system to an actual spatial coordinate system. A liquid level position measuring method using image processing, which is characterized by obtaining an actual position. 5. The scale scale according to claim 1, wherein the side wall is provided with a scale scale and numbers in advance, and the scale scale value is read by image analysis from an image captured by the camera to obtain the scale scale value. A liquid level position measuring method using image processing in which when the reading cannot be performed, it is determined that the scale portion is below the liquid level, and the final scale scale reading position is the actual position of the liquid surface. 6. The method according to claim 1, wherein when the boundary position between the side wall and the liquid upper surface is obtained by image processing, the discolored portion of the boundary portion due to the capillary phenomenon and the reflection image of the discolored portion on the liquid surface are image-analyzed. A liquid level position measuring method using image processing in which an upper boundary line and a lower boundary line of the discolored portion are obtained and a liquid surface position is an intermediate position between the upper boundary line and the lower boundary line thus obtained. 7. The image according to claim 1, wherein when a projection is present in the liquid, an image of the projection above the liquid and a reflection image of the projection above the liquid surface in the image captured by the camera are provided. A liquid surface position measuring method using image processing, wherein the boundary position between the side wall and the liquid upper surface is estimated based on the positional relationship on the image.