JP2000329519A - Coil position detector - Google Patents

Abstract

PROBLEM TO BE SOLVED: To obtain a coil position detector capable of accurately detecting the coil width and the center position even in the case of receiving a strong regular reflection light. SOLUTION: A laser light scanning device 12 linearly scans a laser spot light, forms a plurality of slit lights and irradiates a coil with them and a TV camera 10 images a coil which is irradiated with the laser light. An image synthesizing means 14 respectively synthesizes a brightness image of the laser light and its corresponding cast angle of the laser light. A shape operation means 16 obtains three-dimensional data of the laser irradiation part of the coil from the brightness image of the laser light and the case angle of the laser light. The regular reflection light noise removal means 16 removes the regular reflection light noise among three-dimensional shape data by the shape operation means 15. A coil operation means 17 detects the radial center position of the coil and the coil diameter and detects the width direction center position of the coil and the coil width based on the three dimensional shape data from which the regular reflection light noise is removed.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a coil position detecting device applied to an unmanned operation of an overhead crane in a coil yard of a steelworks.

[0002]

2. Description of the Related Art Conventionally, as this type of coil position detecting device, for example, Japanese Patent Application Nos. 8-349207 to 8-34.
9209, Japanese Patent Application No. 8-349105 to Japanese Patent Application No. 8-3
49107 and Japanese Patent Application No. 9-003299 have been proposed. The measurement system in these applications is:
In each case, the laser spot light of the laser light source attached to the overhead crane is linearly scanned to form a plurality of slit lights and irradiate the coil, and the coil irradiated with the laser light is irradiated by a television camera. The width of the coil and the center position are obtained by taking an image and performing arithmetic processing on the video signal of the television camera.

[0003]

In the above-described coil position detecting device, the reflected light of the laser spot light applied to the coil surface is imaged by a television camera. Since the laser light applied to the coil surface spreads over a wide angle, it is possible to always receive stable reflected light regardless of the shape of the coil surface. However, when the surface property of the target coil is close to a mirror surface, the amount of light that can be received by the television camera greatly changes depending on the position where the laser spot light is applied.

For example, as shown in FIG. 18, when the coil surface 70 is mirror-shaped, when the laser light 72 emitted from the laser light source 71 impinges on the coil surface 70 and is reflected, the laser reflection at the point A occurs. Since the light is reflected in the regular reflection light direction determined by the inclination of the coil surface 70 at the point A (see the reflected light 73a in the figure), almost no light is incident on the television camera 10. In contrast, B
At the point, the inclination of the coil surface at the laser light source 71 and the point B and the television camera 10 have a positional relationship of regular reflection, and when the laser spot light is applied to the point B, the television camera 10 Intense laser light is incident (see reflected light 73b in the figure). FIG. 18 is a diagram showing the positional relationship among the laser light source 71, the coil surface 70, and the television camera 10. Reference numerals 73a and 73b denote reflected light at the time of a mirror surface.
Reference numerals a and 74b denote reflected light on the diffusion surface.

As described above, when the very strong reflected light 73b enters the television camera 10, the reflected light 73b is coupled to the imaging surface due to reflection and scattering inside the optical system composed of the lens of the television camera 10 and the imaging surface. In the image, the laser spot light spreads greatly, and in some cases, the lens and the CCD
Phenomena such as a ghost image generated between the images appear.
For this reason, when the position detection of a coil having a surface property having a very high reflectivity is measured by the above-described measurement method, as shown in FIG. When the positional relationship between the laser light source 71, the coil surface 70, and the television camera 10 is in the regular reflection light receiving state, the laser spot light is greatly spread in the place in the regular reflection light receiving state. . Alternatively, since a ghost image is generated, there is a problem that laser spot light or laser slit light cannot be accurately detected. However, because the coil shape is cylindrical,
The location of the regular reflection state is determined by the arrangement of the laser light source 71 and the television camera 10 and the inclination of the coil surface 70, and is limited to a specific portion of the coil surface 70. FIG. 19 is an explanatory diagram of the binarized image (luminance image) 75.
Reference numeral 6 denotes an image corresponding to the laser slit light (hereinafter, simply referred to as laser slit light), reference numeral 77 denotes a coil to be measured, and reference numeral 78 denotes an image when the laser light is regularly reflected and strong light is returned. .

[0006] In the case of a target material having a coil surface property close to a mirror surface, the above-mentioned phenomenon occurs, and therefore, regular reflection occurs in a conventional coil position detecting device (Japanese Patent Application No. 9-003299). When the location is near the edge portion in the coil width direction, as shown in FIG. 19, the spread of the slit light caused by the specular reflection reaches the outside of the coil edge, and it is as if a portion where the coil does not actually exist. The slit light is detected as if the coil existed. Therefore, when the coil width and the center position are calculated by the conventional coil position detecting device, the coil width is detected to be widened by the amount detected by the laser slit light being spread. Is no longer required.

SUMMARY OF THE INVENTION The present invention has been made to solve the above-mentioned problems, and enables the coil width and the center position to be determined with high accuracy even when the above-mentioned strong regular reflection light is received. It is an object of the present invention to provide a coil position detecting device.

[0008]

(1) A coil position detecting device according to the present invention is mounted on an overhead crane that carries a coil, and scans a laser spot light linearly to form a plurality of slit lights. And irradiate the coil,
A laser beam projecting unit that outputs a projection angle of the laser beam at that time, an image capturing unit that is attached to the overhead crane, and captures an image of a coil irradiated with the laser beam, and an image captured by the image capturing unit and the laser beam. An image synthesizing means for synthesizing a laser light luminance image and a laser light projection angle corresponding to the luminance image of the laser light from the light projection angle of the light projecting means, and a coil from the laser light luminance image and the laser light projection angle. Shape calculating means for obtaining three-dimensional shape data of the laser beam irradiation unit;
Among the three-dimensional shape data by the shape calculating means, the area of the laser slit light at each projection angle of the laser light is obtained, and when the area is equal to or larger than a predetermined value, the data at the projection angle at that time is removed. The regular reflection light noise removing means and the three-dimensional shape data from the regular reflection light noise removing means are projected in the width direction of the coil to obtain the radial center position and the coil diameter of the coil, and the 3 And a coil calculating means for projecting the dimensional shape data in the radial direction of the coil parallel to the reference plane and obtaining a center position and a coil width in the width direction of the coil.

(2) The coil position detecting device according to the present invention removes specular reflection light noise before obtaining three-dimensional shape data. In the combined image by the combining means, the area of the laser slit light at each projection angle of the laser light is obtained,
When the area is equal to or larger than a predetermined value, a regular reflection light noise removing means for removing data based on the two-dimensional projection angle at that time is provided.

(3) In the coil position detecting device according to the present invention, when the intensity of the laser slit light exceeds a predetermined threshold value instead of the regular reflection light noise elimination means of (1) and (2). , Which removes data according to the projection angle at that time.
And a regular reflection light noise removing means.

In the present invention, as described above, the specular reflected light noise removing means detects and removes strong regular reflected light, thereby eliminating the limitation on the surface properties of the coil to be measured, and providing an accurate coil. The center position in the radial direction, the coil diameter, the center position in the width direction of the coil, and the coil width can be obtained.

[0012]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS FIG. 1 is a configuration diagram for explaining a measurement principle of a coil position detecting device according to the present invention. In the figure, a DUT 2 is placed on a reference plane (floor surface) 1, a television camera 10 is arranged on the reference plane 1 such that the central axis at the time of photographing is orthogonal, and The slit light scanning device 12 is arranged at a position different from the camera 10. The slit light scanning device 12 has a slit light 12a
And scan it. The television camera 10 captures an image of the state, and the image signal is input to the image synthesizing device 14 together with the projection angle θ from the slit light scanning device 12, and the image signal is processed to generate a synthesized image. The shape calculation device 15 determines the position of the DUT 2 based on the synthesized image.
Find the dimensional shape. The regular reflection light noise elimination device 16 removes a noise component due to strong regular reflection light of the slit light, which becomes a noise when obtaining the position, width, and the like of the coil. The coil calculation device 17 obtains the position, width, and the like of the coil as the device under test 2 based on the three-dimensional shape from which the noise component has been removed.

Next, the shape measuring apparatus shown in FIG. 1 will be described in terms of a measuring method, an image synthesizing method, information on the position and projection angle of laser slit light, and generation of a shape calculation principle.

(A) Measuring method The optical system uses an optical system generally called a light section method. FIG. 2 is a diagram showing a configuration of the slit light scanning device 12. The slit light scanning device 12 first generates a slit light by scanning a laser spot light from the semiconductor laser 20 with a first scanning mirror 22 whose rotation is controlled by a drive motor 22a. At this time, the direction of the laser slit light is set to be in the y-axis direction in FIG.
Then, the projection position of the laser slit light is controlled in the x-axis direction in FIG. 1 by the second scanning mirror 24 whose rotation is controlled by the drive motor 24a. At the same time, the projection angle θ of the scanning mirror 24 at this time is input to the image synthesizing device 14 as an angle signal. Then, the irradiation pattern of the laser slit light 12 a is imaged by the television camera 10, and the video signal is input to the image synthesizing device 14.

The image synthesizing device 14 performs image processing on a video signal of the television camera 10, and generates a composite image from the video signal and the laser projection angle θ at that time. Then, the shape calculation device 16 calculates the three-dimensional shape of the laser beam irradiation section of the device under test 2 based on the principle of triangulation described later using the synthesized image. Next, the projection angle of the second scanning mirror 24 is rotated from θ by a small angle Δθ, and the first scanning mirror 22 projects the laser slit light in the y-axis direction again. In this way, the shapes of a plurality of cross sections are measured one after another while changing the projection angle of the laser slit light, and the overall shape of the DUT 2 is reconstructed by synthesizing the respective light cross sectional shapes. The rotation pitch Δθ of the second scanning mirror 24
The finer the resolution, the more the shape resolution of the DUT 2 improves.
However, on the other hand, the measurement time becomes longer. Further, the slit light scanning device 12 of FIG. 2 is provided with a shielding plate 26, which is controlled in synchronization with the drive motor 24a.
While the projection angle θ of the laser slit light is changing, the laser slit light is not irradiated on the DUT 2. Alternatively, the semiconductor laser 20 may be controlled so that no laser light is generated during that time. In any case, unnecessary light is prevented from being taken in between the light section line to be measured.

(B) Image Synthesizing Method FIG. 3 is a block diagram showing the configuration of the image synthesizing device 14. The video signal from the television camera 10 is converted into a digital signal by the A / D converter 30 and input to the maximum luminance image calculation unit 31. This maximum brightness image calculation unit 31
It is composed of a comparator 31a, a selector 31b and a maximum luminance image memory 32. The maximum luminance signal of each pixel is obtained and stored in a corresponding pixel area of the maximum luminance image memory 32. That is, when the video signal _ac is input via the A / D converter 30, the comparator 31a outputs the signal a of the pixel corresponding to the video signal in the maximum luminance image memory 32.
reads _m, compared with the luminance level of the signal a _m brightness level and a maximum brightness image memory 32 of the video signal a _c,
When a _c > a, a selector signal is sent to the selectors 31b and 37a. When the selector 31b inputs the selector signal, the maximum brightness image memory 32 in the video signal a _c
Rewrite signal a _m. By performing such processing in order for the input video signal, the value of each pixel area of the maximum luminance image memory 32 becomes the maximum luminance level.

A video signal from the television camera 10 is converted into a digital signal by an A / D converter 33 and input to a minimum luminance image calculation unit 34. The minimum luminance image calculation unit 34 includes a comparator 34a, a selector 34b, and a minimum luminance image memory 35. The minimum luminance signal of each pixel is obtained and stored in a corresponding pixel area of the minimum luminance image memory 35. That is, the comparator 34a
Is, A / the via D converter 33 to the input video signal a _c, reads out the signal a _m of the pixels corresponding to the video signal of the minimum brightness image memory 35, a video signal a _c luminance level and minimum luminance image It compares the luminance level of the signal a _m of the memory 35, and sends a selector signal to the selector 34b in the case of a _c <a _m. The selector 34b is by entering the selector signal, it rewrites the signal a _m of the minimum brightness image memory 35 in the video signal a _c. By performing such processing in order on the input video signal, the value of each pixel area of the minimum luminance image memory 35 becomes the minimum luminance level.

Also, the projection angle signal θ of the laser slit light
Are converted into digital signals by the A / D converter 36 and input to the coded image operation unit 37. The coded image operation unit 37 includes a selector 37a and a projection angle image memory 3
8, and obtains a projection angle signal θ of the laser slit light when the maximum luminance signal is obtained for each pixel, and determines the projection angle signal θ in the pixel area of the projection angle image memory 35 corresponding to the maximum luminance image memory 32. Store. That is, when the comparator 31a outputs a selector signal and the selector 37a receives the selector signal, the light emission angle signal b at that time is output.
rewriting the signal b _m of the projection angle image memory 38 _c. In By performing such processing in order on the input video signal, the projection angle signal image memory 38 stores the projection angle signal θ when each pixel has the maximum luminance level.

FIG. 4 is an explanatory diagram conceptually showing the arithmetic processing of the image synthesizing device 14. As described above, the image synthesizing device 14 performs real-time processing on the video signal captured by the television camera 10 to detect the luminance level of each pixel,
The maximum luminance image calculation is performed for each pixel, and the luminance signal at the timing when the luminance level of each pixel becomes the highest (that is, when the laser beam hits the corresponding point in the visual field) is stored in the maximum luminance image memory 32. At the same time, the projection angle information θ at that timing is recorded in a corresponding pixel of the projection angle image memory 38 separately provided in addition to the luminance image,
It simultaneously has a luminance image in which a plurality of laser slit lights are combined and projection angle information θ of each slit light.

The extraction of the portion irradiated with the laser slit light may be performed by performing a binarization process from the maximum luminance image to extract a point (ie, a slit light portion) at which the laser light is irradiated and the luminance becomes high. In the present embodiment, as described above, the minimum luminance image calculation is simultaneously performed and stored in the minimum luminance image memory 35, whereby the calculation of (maximum luminance image−minimum luminance image) is performed and the calculated result is stored in the luminance image memory 42. , The luminance level of the background light is removed, and only the component whose luminance level is increased by the irradiation of the laser light is extracted and binarized. By extracting the area (light cutting line) irradiated with the laser slit light from the luminance memory 42 by this processing, the light emitting angle information of each light cutting line is identified from the light emitting angle images of the light emitting angle image memories 38 and 41. I do.

(C) Extraction of the position of the laser slit light and the projection angle information The image synthesizing device 14 calculates the position of the slit light from the maximum luminance image, the minimum luminance image and the projection angle image synthesized as described above. The following image operation is performed to extract the luminance level of the background light. Luminance image = Maximum luminance image−Minimum luminance image (1) By the image calculation of the above equation (1), a part of the luminance that has become brighter due to the irradiation of the laser beam is obtained. The area subjected to the laser light is extracted by the value processing.

(D) Principle of Shape Calculation The shape calculation device 15 calculates the three-dimensional coordinates of the laser cut light of the multi-slit light generated by the slit light scanning device 12 to obtain the entire shape of the DUT 2. . This shape calculation is based on a triangulation method. Laser slit light 12a is projected from the slit light scanning device 12 onto the DUT 2 at a projection angle θ, and the state is imaged by the television camera 10. The height z (x ′, y ′) of the point (x ′, y ′) at which the laser beam extracted from the image captured by the television camera 10 hits is given by the following equation in consideration of the perspective effect of the television camera 10. The relationship holds. z (x ′, y ′) = Z ₀ − {X ₀ − (1−z (x ′, y ′) / a) x} tan θ (x ′ · y ′) (2) x ′: TV camera position y on the reference plane of the captured image ': position on the reference plane of the image captured by the television camera X _0: x-coordinate Z ₀ of the laser beam scanning rotation axis: z coordinates of the laser beam scanning rotation axis a: distance between the television camera and the reference plane θ: laser slit light projection angle If the above equation (2) is modified, the height z (x ′, y ′) of the point where the laser slit light hits is given by the following equation. Is determined by

[0023]

(Equation 1)

The coordinates (x, y) on the three-dimensional coordinates of the point (x, y) where the laser beam is extracted on the image are given by the following equations. x = ｛1−z (x ′, y ′) / a｝ x ′ (4) y = ｛1−z (x ′, y ′) / a｝ y ′ (5) It is clear from the above description. As described above, the coordinates x ', y' on the image of the television camera 10 at the projection angle .theta.
It can be seen that the three-dimensional coordinates x, y, z of the detected point are obtained by the above equations (3), (4), (5). The above equations (4) and (5) correct the correction for the perspective effect that the closer to the TV camera 10 the larger the closer to the TV camera 10 because the distance between the DUT 2 and the TV camera 10 is finite. It can be seen that accurate shape detection cannot be performed only with a two-dimensional image.

Now that the principle of shape measurement in the present invention has been clarified by the above description, an example in which the coil position detector of FIG. 1 is applied to the detection of a yard coil will be described. FIGS. 5A, 5B, and 5C are a plan view, a front view, and a view of the television camera 10 when the optical thread (the television camera 10 and the slit light scanning device 12) of FIG. FIG. As shown in FIG. 2A, the television camera 10 is disposed on the trolley 43 in a direction rotated by 45 ° with respect to the direction of its traverse, and the optical axis thereof is positioned vertically downward. The slit light scanning device 12 is installed on the x-axis of the television camera 10 and the direction of the slit light is
It is adjusted so as to be parallel to the y-axis direction of 0. The coil 3 to be measured is placed parallel to the crane's traversing direction or running direction. Therefore, the attitude of the coil 3 captured by the television camera 10 is 45 ° in the camera view 44.
It is in a tilted (rotated) state. In the yard, the coil 3 is arranged in either the traverse direction or the traveling direction of the crane, and the arrangement direction is grasped by a system computer managing the yard.

In this state, the slit light scanning device 12 irradiates the upper surface of the coil 3 while scanning the laser slit light,
The television camera 10 obtains a slit light image obliquely at 45 degrees to the coil 3, the image synthesizing device 14 generates a synthetic image by the above-described method, and the shape calculating device 15 generates a laser beam based on the synthetic image. The height of the hit point is calculated based on the above calculation formula. By sequentially performing this operation for a plurality of lines, it is possible to obtain a plurality of cross-sectional shapes of an object in a measurement visual field such as a coil / cart and a floor surface.

Next, the coil calculating device 17 uses the shape data to identify the coil and the peripheral portion (the floor and the truck) from the difference in height, since the coil is higher than the floor and the truck. Before extracting only the shape of the coil, the following processing is performed by the specular reflection light noise elimination device 16 to remove noise due to the specular reflection light.

The regular reflection light noise elimination device 16 uses the light projection angle image θ and the shape calculation image obtained by the shape calculation means 15 to calculate a noise from a measurement target coil which becomes noise when calculating the coil shape and position. Is detected and strong specular reflection data is removed. Specific processing contents will be described below.

When a coil in which strong regular reflection light does not return is measured, a binarized image and a projection angle image as shown in FIGS. 6 and 7 are obtained. When a coil that returns strong regular reflection light is measured, a binarized image and a projection angle image as shown in FIGS. 8 and 9 are obtained. As described above, each laser slit light is coded at one projection angle, and a code value is stored in each pixel to which the laser slit light is applied. In the examples shown in FIGS. 8 and 9, strong regular reflection light is generated at the projection angle θ = θ4.

Since the projection angle images of FIGS. 7 and 9 are coded for each projection angle θ of the laser slit light as described above, a histogram in which the number of pixels at each projection angle θ is counted is used as a histogram. When it is made, it will be as shown in FIG. 10 and FIG. When measuring a coil that does not return strong specular reflection light, the number of pixels at each projection angle does not change significantly (FIG. 10), but when measuring a coil that returns strong specular reflection light, specular reflection occurs. The number of pixels at each light projection angle at which occurs increases due to the spread of light (FIG. 11). In the example of FIG. 11, at the light projection angle θ4, the value is high because a plurality of images are coded at the same light projection angle θ4 because the laser spot light is spread and picked up due to strong regular reflection light. Will be. From this, it is possible to determine whether or not an image in which strong regular reflection light has been obtained is obtained by creating a histogram for each projection angle and examining the number of pixels. As shown in FIG. 12, a threshold is set for the histogram,
When the number of pixels exceeds that number, it can be determined that there is strong regular reflection light. If the shape calculation image obtained by extracting only the shape of the coil is processed by the coil calculation device 17 with the data containing strong specular reflection light, the spread of the spot light becomes noise and the correct shape cannot be obtained. It must be excluded from the data (see FIGS. 13, 14, and 15).

Therefore, it is necessary to obtain the coordinates of the pixel having the code of the determined projection angle θ having strong regular reflection light from the projection angle image, and to remove the pixel data of the same coordinate in the shape calculation image. Specifically, as shown in FIG.
A threshold value is set in the histogram, and if the number of pixels exceeds the threshold value, it is determined that there is strong regular reflection light, and the coordinates of the pixel are obtained. Then, regular reflection noise is removed by removing the data of the projection angle θ4 from the shape calculation image (see FIG. 13). The coil calculating means 17 obtains a cross-sectional shape of the coil in the radial direction and the width direction in order to detect the position of the coil from the noise-removed shape calculation image.

FIG. 16 is an explanatory diagram when the coil calculating means 17 determines the cross-sectional shape of the coil in the radial direction and the width direction. Here, the shape data is projected in a direction in which the shape data of the coil 3 is rotated by an installation angle (45 °) of the television camera 10 as a detection head, that is, in a traversing direction of the trolley 43 and a traveling direction. As a method of projection,
In the coil width direction, an average value of the coil width direction shape data is obtained. In the illustrated example, for example, an average value of the slit lights 51, 52, and 53 at, for example, the illustrated points is obtained, and the average value is used as the value of the point in the projected shape in the coil width direction. , 53, and 54, the average value is also obtained, thereby obtaining the projected shape (shape data) 55 in the coil width direction. Further, by taking the maximum value of the shape data in the coil radial direction in the coil radial direction, accurate shape data can be obtained. In the illustrated example, the projection shape (shape data) 56 in the coil rear direction is obtained by taking the maximum value of the data 51a, 52a, 53a, 54a obtained by projecting the slit light beams 51, 52, 53, 54 in the coil radial direction. Ask.

Although the above description is an example in the case where the arrangement direction of the coil 3 is known in advance, the same processing can be performed when the direction is not known in advance. In such a case, the direction is temporarily determined, and then the projected shape (shape data) in the coil width direction and the radial direction is obtained, and the arrangement direction can be determined from the shape data. If the direction is incorrect, the projection shape (shape data) in the coil radial direction is obtained for the determined arrangement direction.

Next, a method for obtaining the coil width direction position, the coil width, the coil radial direction center position, and the coil diameter from the shape data projected in the coil width direction and the radial direction will be described. Regarding the coil width direction, as shown in FIG. 17, by detecting points at both left and right ends exceeding height data set separately from the projected shape data in the width direction, the points are shifted to the right and left of the coil. Recognize as an edge. Therefore, the center position and the width in the coil width direction can be obtained by the following equations. Coil center position = (right edge + left edge) / 2 (6) Coil width = | right edge−left edge | (7)

When the above processing is performed, FIGS.
As shown in the figure, if there is spot light data spread by strong regular reflection light, the coil width is determined to be wide and the position of the coil is not measured correctly. However, in the present embodiment, as described above, the strong regular reflection light There is no such adverse effect because the data is removed. However, in order to remove specularly reflected light noise, as shown in FIG.
B and C), at least one slit light data needs to be applied to the coil edge in the width direction after noise removal. Therefore, it is necessary to determine the number of slit lights (light projection angle intervals) such that two or more slits are applied to the width direction coil edge even with the smallest coil diameter to be measured.

In the coil radial direction, since the projected coil shape becomes a semicircle, the center position and the diameter of the coil can be obtained by substituting the shape data into the equation of the circle. (X−c _x ) ² + (z−c ₂ ) ² = r ² (8) x, y: semi-circular shape data of the coil obtained by projection, c _x , c ₂ : center position of coil rear direction r : Coil radius

Specifically, if there is data at three or more points in the radial direction of the coil, the above equation (8) can be solved. However, the measurement data in the radial direction of the coil can be used effectively to improve the reliability of the measurement. In order to improve the accuracy and accuracy of the measurement data in the coil radial direction, a plurality of combinations of three points are set, noise is removed from a plurality of center positions obtained by each calculation, and then statistical processing is performed. A highly reliable coil center position can be obtained.

As a method of removing noise, since noise takes an abnormal value (a value abnormally larger or smaller than the true value) with respect to the true value, the distribution of measured values is obtained and set in advance by majority rule theory. There is a method of removing points apart from the set range, and a method of deleting data of points set in advance in order of increasing distance from the average value of all the center positions obtained by the three-point calculation. Such a process is naturally used also at the stage of obtaining the cross-sectional shape in the coil radial direction and for removing abnormal shape data from the obtained cross-sectional shape.

As a method for obtaining the true center position from a number of coil center positions, a method for obtaining an average value of many points, and a method of dividing a distribution of many points into a mesh and determining a position at which the maximum value of the distribution is obtained. There is a method to ask.

In this embodiment, as is apparent from the above description, even if there is strong specular reflection light from the coil to be measured, which becomes noise when measuring the coil shape and position, there is strong specular reflection light. By removing the strong specular light data that becomes noise, eliminating the limitation of the surface properties of the coil to be measured,
It is possible to accurately determine the radial center position and coil diameter of the coil, and the center position and coil width of the coil in the width direction.

In the present embodiment, after performing the shape calculation based on the synthesized image of the laser slit light, the data of the slit light that has caused noise due to the regular reflection is removed.
Before performing the shape calculation, the above processing may be performed to remove noise. Also, in removing noise,
When the signal level of the laser slit light (image) exceeds a predetermined threshold level instead of obtaining the area of the laser slit light based on each two-dimensional projection angle of the laser light,
The data based on the two-dimensional light projection angle at that time may be removed.

[0042]

As described above, according to the present invention, in measuring the shape and position of the coil, even if there is strong specular reflection light from the coil to be measured, which is noise, there is strong specular reflection light. Detection and removal of strong specular reflected light data, which is noise, eliminates restrictions on the surface properties of the coil to be measured, and allows accurate coil radial center position, coil diameter, and coil width center position. And the coil width can be obtained.

[Brief description of the drawings]

FIG. 1 is a configuration diagram for explaining a measurement principle of a coil position detection device of the present invention.

FIG. 2 is a diagram illustrating a configuration of a slit light scanning device in FIG. 1;

FIG. 3 is a block diagram illustrating a configuration of the image synthesizing apparatus in FIG. 1;

FIG. 4 is an explanatory diagram conceptually showing a calculation process of the image synthesizing device of FIG. 1;

5A and 5B are explanatory diagrams of a case where the coil position detecting device of FIG. 1 is applied to coil position detection of a yard, where FIG. 5A is a plan view, FIG. 5B is a front view, and FIG. FIG.

FIG. 6 is an explanatory diagram of a binarized image when there is no strong regular reflection light.

FIG. 7 is an explanatory diagram of a projection angle image θ when there is no strong regular reflection light.

FIG. 8 is an explanatory diagram of a binarized image when there is strong regular reflection light.

FIG. 9 is an explanatory diagram of a projection angle image θ when there is strong regular reflection light.

FIG. 10 is a histogram of a two-dimensional light projection angle when there is no strong regular reflection light.

FIG. 11 is a histogram of a two-dimensional light projection angle when there is strong regular reflection light.

12 is an explanatory diagram for setting a threshold value in the histogram of FIG. 11 and detecting a projection angle of strong regular reflection light.

FIG. 13 is a projection angle image when strong regular reflection light is present,
FIG. 4 is an explanatory diagram showing a relationship between a shape calculation image (before regular reflection noise removal) and a shape calculation image (after regular reflection noise removal).

FIG. 14 is an explanatory diagram in a case where a cross-sectional shape in a radial direction and a width direction of a coil is obtained in a state where specular reflection noise is included.

FIG. 15 is an explanatory diagram for determining a cross-sectional shape in a radial direction and a width direction of a coil after removing regular reflection noise.

FIG. 16 is an explanatory diagram for determining a cross-sectional shape in a radial direction and a width direction of a coil.

FIG. 17 is an explanatory diagram for obtaining a coil width or the like from shape data in the width direction projected in the coil width direction.

FIG. 18 is a diagram showing a positional relationship between a laser light source, a coil surface, and a television camera.

FIG. 19 is an explanatory diagram of a binarized image (luminance image) when there is strong regular reflection light.

 ────────────────────────────────────────────────── ─── Continuing from the front page (72) Inventor Mitsuaki Uesugi 1-2-1, Marunouchi, Chiyoda-ku, Tokyo F-term in Nihon Kokan Co., Ltd. (reference) 2F065 AA04 AA22 AA53 BB05 CC06 DD03 DD04 FF09 GG06 GG12 HH06 JJ03 JJ26 LL13 MM16 QQ03 QQ06 QQ21 QQ24 QQ34 QQ42 QQ43

Claims (4)

[Claims] 1. An overhead crane for transporting a coil, linearly scans a laser spot light to form a plurality of slit lights, irradiates the coil, and a projection angle of the laser light at that time. Laser beam emitting means for outputting a laser beam; imaging means attached to the overhead crane for imaging a coil irradiated with the laser light; and an image captured by the imaging means and projection of the laser light emitting means. Image synthesizing means for synthesizing the brightness image of the laser beam and the projection angle of the laser beam corresponding thereto from the light angle, and the laser beam of the coil from the brightness image of the laser beam and the projection angle of the laser beam A shape calculating means for obtaining three-dimensional shape data of the irradiating part; First, when the area is equal to or larger than a predetermined value, a regular reflection light noise removing unit that removes data according to the projection angle at that time; And the coil radial position and the coil diameter are obtained, and the three-dimensional shape data of the laser beam irradiating section of the coil is projected in the coil radial direction parallel to the reference plane, and the coil width direction central position is obtained. And a coil calculating means for calculating a coil width. 2. A laser beam, which is attached to an overhead crane for transporting a coil, linearly scans a laser spot light to form a plurality of slit lights, irradiates the coil, and projects a projection angle of the laser light at that time. Laser beam emitting means for outputting a laser beam; imaging means attached to the overhead crane for imaging a coil irradiated with the laser light; and an image captured by the imaging means and projection of the laser light emitting means. Image synthesizing means for synthesizing the luminance image of the laser light and the projection angle of the laser light corresponding thereto from the light angle, and a laser slit light based on the projection angle of the laser light among the synthesized images by the image synthesis means The regular reflection light noise removing means for removing data according to the projection angle at that time when the area of the regular reflection light noise is obtained. Shape calculating means for obtaining three-dimensional shape data of a laser light irradiating portion of a coil from a luminance image of the laser light from the removing means and a projection angle of the laser light; and projecting the three-dimensional shape data in a width direction of the coil. The center position and the coil diameter of the coil in the radial direction are obtained, and the three-dimensional shape data of the laser beam irradiating section of the coil is projected in the radial direction of the coil parallel to the reference plane, and the center position and the coil width of the coil in the width direction are obtained. A coil position detecting device comprising: a coil calculating means for obtaining the following. 3. The method according to claim 2, wherein the specular reflected light noise removing means is replaced with:
3. The apparatus according to claim 1, further comprising a second regular reflection light noise removing unit that removes data according to the projection angle when the intensity of the laser slit light exceeds a predetermined threshold. Coil position detection device. 4. The image synthesizing unit sequentially captures and stores signals of respective pixels in a screen of an image captured by the imaging unit, and stores the luminance level of a signal input later for the same pixel. The luminance level of the signal being compared is compared, and when the luminance level of the signal input later is higher, the storage content of the pixel is updated with the luminance level, and the maximum luminance level for each pixel is obtained. A brightness image calculation unit, and for each pixel in the screen of the image, when the maximum brightness image calculation unit updates the storage content of the pixel, generates a composite image using the projection angle information at that time as the value of the pixel A coded image calculation unit that sequentially captures and stores signals of respective pixels in a screen of an image captured by the imaging unit, and stores a luminance level of a signal input later for the same pixel. And compares the luminance level of the already stored signal with the luminance level of the already stored signal. A minimum brightness image calculation unit for obtaining a brightness level, and for each pixel, a difference between the maximum brightness level and the minimum brightness level is extracted to extract a rise in the brightness level due to laser light irradiation, and a laser slit light irradiation area. The coil position detecting device according to any one of claims 1 to 3, wherein