JPH10105719A - Optical measurement method for hole position - Google Patents

Abstract

PROBLEM TO BE SOLVED: To accurately measure the position of a hole picture appearing in an image picked up from a hole of a work. SOLUTION: A differentiated picture where a hole edge part (br) of the hole picture is generated by differential processing of a picture with variable density (A), and a template TP is used for the differentiated picture to obtain the position of an approximate center m' of the hole picture by pattern matching (B). Scanning lines LSC are set in plural radial directions with the approximate center m' as the center, and coordinates of a hole edge point Pb coinciding with each scanning line in the hole edge part of the hole picture are detected in accordance with the luminance distribution on this scanning line (C).

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a method for optically measuring the position of a hole provided in a work.

[0002]

2. Description of the Related Art Conventionally, as a method of this kind, a first step of determining a rough center position of a hole image appearing on an image pickup screen when a workpiece is picked up, and a hole image based on the rough center position determined. A second step of setting a plurality of measurement points so as to intersect a plurality of points of the hole edge of the hole image, and a third step of detecting coordinates of a hole edge point matching each measurement point of the hole edge of the hole image. A method is known in which a regression circle equation approximated to a hole image is obtained from the coordinates of the hole edge points by regression processing, and the center coordinates of the regression circle are used as the center coordinates of the hole image.

In this apparatus, a binarized screen is created by binarizing an imaged screen with shading obtained when a workpiece is imaged, and the area center of gravity of a hole image serving as a dark portion in the binarized screen is formed. In the second step, a plurality of vertical and horizontal windows, which are measurement locations on the binarized screen with reference to the center of gravity, are applied to the hole edge of the hole image in the second step. In the third step, the coordinates are detected with the light-dark boundary in each window as the hole edge point.

[0004]

In the above conventional example,
A window set at a position vertically or horizontally away from the center of gravity becomes oblique to the hole edge of the hole image, and the light-dark boundary in the window extends obliquely, and the coordinates of the hole edge point Is difficult to detect accurately.

SUMMARY OF THE INVENTION In view of the foregoing, it is an object of the present invention to provide a method capable of accurately detecting the coordinates of a hole edge point and improving the measurement accuracy of a hole position.

[0006]

Means for Solving the Problems In order to solve the above problems,
The present invention is a method for optically measuring the position of a hole provided in a work, comprising: a first step of calculating an approximate center position of a hole image appearing on an imaging screen when the work is imaged; Setting a plurality of measurement locations so as to intersect a plurality of locations on the hole edge of the hole image based on the approximate center position
And a third step of detecting coordinates of a hole edge point that matches each measurement point of the hole edge of the hole image, wherein the second step includes a plurality of radiations centered on the approximate center position. The apparatus is configured to set the plurality of measurement points along a direction.

According to the present invention, each measurement point can be set so as to be substantially perpendicular to the edge of the hole in the light-section image. for that reason,
The detection accuracy of the coordinates of the hole edge at each measurement location can be improved, the measurement accuracy of the center coordinates of the hole image obtained from the coordinates of these hole edges can be improved, and the hole position can be accurately measured.

[0008] A binarized screen is created by binarizing the imaged screen with shading, and a long window in the radial direction is set as the measurement location on the binarized screen. It is also possible to detect the coordinates of the portion as the hole edge point, but an image portion due to disturbance light (for example, light reflected from the hole inner surface) may appear as a bright portion on the binarized screen, The light-dark boundary in the window is shifted from the normal hole edge of the hole image, and a measurement error is likely to occur.

On the other hand, each of the measurement points is set as a scanning line extending in each of the radiation directions over a predetermined range determined based on a hole diameter stored in advance, and
In the process, if the coordinates of each hole edge are detected from the brightness distribution on each scanning line of the imaged screen with shading, the influence of disturbance light is eliminated as much as possible, and the coordinates of each hole edge are changed. It can be detected with high accuracy, which is advantageous.

When the coordinates of the hole edge point are detected from the imaging screen with shading as described above, the approximate center position of the hole image can be determined without binarizing the screen. It is desired. Here, when a differentiated screen is created by differentiating the luminance distribution of the imaging screen with shading, the differentiated screen is a screen that emphasizes the edge of the hole of the hole image. Therefore, if pattern matching is performed on the differentiated screen using a template representing the shape of the hole edge of the hole image, the approximate center position of the hole image can be determined, and the above-described demand is met.

[0011]

FIG. 1 shows a work A such as an automobile body.
1 shows an outline of an optical measuring device used for measurement of the object A. The device includes a slit light source 1 composed of a slit laser or the like for irradiating a workpiece A with a slit light, an imager 2 composed of a CCD camera, and an imager 2 composed of a CCD camera. It comprises a spot light source 3 composed of a group of light emitting diodes arranged in a ring around the lens 2a, and an image processing circuit 4 for inputting an image signal from the image pickup device 2.

The slit light source 1, the image pickup device 2, and the spot light source 3 are mounted on a measuring head (not shown) attached to an operating end of a moving mechanism such as a robot, and the measuring head is located at a position facing a plurality of measuring parts of the work A. Move in order to measure each measurement site. The slit light source 1 and the image pickup device 2 are mounted on the measuring head in such a positional relationship that the optical axis of the image pickup device 2 obliquely intersects the optical surface SP of the slit light at a predetermined angle θ (for example, 45 °).

FIG. 1 shows a state in which a measuring head is moved to a position facing a hole B provided in a work A to perform hole measurement. At the time of hole measurement, the work A is first irradiated with a spot light from the spot light source 3 in a state where the image pickup device 2 is directly opposed to the work A. (With shading) is transmitted to and stored in the image processing circuit 4, and then the work A is irradiated with the slit light from the slit light source 1 while the spot light source 3 is turned off. Then, the image data (with shading) is transmitted and stored in the image processing circuit 4.

When the spot light is radiated, the image is displayed on the imaging screen as shown in FIG.
As shown in (A), the hole image b appears as a dark portion, and a light cut image S corresponding to the cross-sectional shape of the work A drawn by the slit light on the surface of the work A at the time of the irradiation of the slit light is captured. As shown in FIG. 3 (B), the light-section image s appears as a bright portion on the imaging screen. Incidentally, when the slit light is applied so as to straddle the hole B, the light cut image s becomes the hole B
Is divided at the part corresponding to.

By the way, the point of intersection of the optical axis of the imaging device 2 and the slit optical surface SP is the origin 0, the optical axis of the imaging device 2 is the Z axis, and the coordinate axis parallel to the slit optical surface SP orthogonal to the Z axis is the Y axis. Consider a spatial coordinate system in which a coordinate axis orthogonal to the Y axis and the Z axis is the X axis. Assuming that a projected image of the spatial coordinate system on the XY coordinate plane is taken by the image pickup device 2, the screen of the image pickup device 2 When the horizontal x-axis corresponding to the X-axis and the vertical y-axis corresponding to the Y-axis are taken with the center point corresponding to the origin 0 as the origin, the x-axis coordinate values and the y-axis coordinate values of the screen are spatial coordinates It represents the horizontal distance and vertical distance from the origin 0 on the XY coordinate plane of the system.
Then, as shown in FIG. 2, the ratio between the x, y coordinates mx, my on the screen of the center m of the hole image b and the X, Y coordinates MX, MY in the spatial coordinate system of the center M of the hole B is obtained from the image pickup device 2. Origin 0
Is equal to the ratio of the distance L from the imager 2 to the work A, and therefore, the displacement amount of the work A in the Z-axis direction is d.
As Z, MX = mx · (L−dZ) / L, and MY = my · (L−dZ) / L, and Z coordinate MZ in the spatial coordinate system of the center M of the hole B
Becomes MZ = dZ.

Here, when the work A is displaced in the Z-axis direction, the slit light plane SP is parallel to the Y-axis and obliquely intersects with the Z-axis, so that the light cut image s is displaced in the x-axis direction on the screen.
The relationship between the x coordinate sx of the light section image s and the X coordinate SX of the slit light irradiating section S on the workpiece A is SX = sx · (L−dZ) / L (1) SX = dZ · tan θ (2) where θ is the inclination angle of the slit optical surface SP with respect to the Z axis. From equations (1) and (2), sx · (L−dZ) / L = dZ · tan θ (3) When the equation (3) is summarized for dZ, dZ = sx · L / (Ltan θ + sx) (4) Thus, if the x-coordinate sx of the light-section image s is measured, the Z-axis displacement dZ of the work A can be calculated from the equation (4), and the spatial coordinates from the center coordinates mx, my and dZ of the hole image b on the screen are obtained. The center positions MX, MY, MZ of the holes B in the system can be determined.

Next, a method of obtaining the center coordinates mx and my of the hole image b will be described. First, a differentiated screen is created by differentiating the brightness distribution of the imaging screen with shading. Since the luminance distribution of the imaging screen changes sharply at the edge of the hole in the hole image b, the differential value increases at the edge of the hole, and the differentiated screen corresponds to the edge of the hole as shown in FIG. A ring-shaped image br appears. Next, a template TP representing the shape of the hole edge of the hole image b is created by graphic processing based on the hole diameter data stored at the time of teaching for measuring the master work, and FIG.
As shown in (B), pattern matching is performed on the differentiated screen by the normalized correlation method or the like, and the position of the approximate center m 'of the hole image b is determined. In order to reduce the processing time of the pattern matching, the pattern matching is performed by reducing the differentiated screen and the template TP at the same ratio (for example, 1/4). By the way, it is possible to binarize the imaging screen with shading and make the area center of gravity of the dark part of the binarized screen the approximate center of the hole image b. However, the dark area of the binarized screen is reflected by light reflected from the inner surface of the hole. May be deformed. In this case, since the area center of gravity of the dark part is greatly deviated from the normal center of the hole image, it is more accurate to determine the position of the approximate center m 'of the hole image b by pattern matching with the differentiated screen as described above.

Next, as shown in FIG. 4C, a plurality of scanning lines LSC, which are measurement locations along a plurality of radiation directions centering on the approximate center m ', are set on the imaging screen with shading. Each scanning line LSC is set to extend over a predetermined range inside and outside the hole edge of the hole image b based on the hole diameter data stored at the time of teaching. And each scanning line L
Based on the luminance distribution on the SC, the coordinates of the hole edge point Pb that matches each of the scanning lines LSC at the hole edge of the hole image b are detected.

FIG. 5A shows a luminance distribution curve on the scanning line LSC, and the luminance sharply decreases at the edge of the hole. Therefore, when a differential curve obtained by differentiating the luminance distribution curve is created, a peak appears at a position corresponding to the edge of the hole as shown in FIG. In this case, the position of the vertex of the differential curve may be set as the position of the hole edge point Pb, but it is difficult to uniquely specify the vertex, and there is variation. Therefore, in the present embodiment, the vertices of the differential curve are determined temporarily, and the points on the differential curve at each of two pixels before and after the pixel (pixel) that matches this vertex are determined. , The equation of the parabola LP approximating the vertex of the differential curve is calculated by regression processing, and the position of the vertex of the parabola is defined as the position of the hole edge point Pb.

Incidentally, the above-described scanning line LS in the radial direction is used.
Instead of C, a scanning line parallel to the x-axis direction and a scanning line parallel to the y-axis direction are formed on both sides in the y-axis direction at the approximate center of the hole image b so as to intersect the hole edge of the hole image b. It is also conceivable to set both sides in the x-axis direction. However, in this case, the scanning line obliquely intersects the edge of the hole, the luminance distribution on the scanning line changes slowly, and the peak of the differential curve becomes low and gentle. Therefore, even when the approximate parabola LP is obtained as described above, the positions of the vertices are apt to fluctuate, and it becomes difficult to obtain the detection accuracy of the hole edge point Pb. On the other hand, if the scanning line LSC is set in the radial direction as in the present embodiment, the scanning line LSC becomes substantially orthogonal to the edge of the hole, and the luminance distribution on the scanning line LSC changes suddenly. The peak of the differential curve is high and steep, and the detection accuracy of the hole edge Pb is improved.

It is also conceivable to binarize a captured image with shading, set a long window in the radial direction as a measurement point on the binarized screen, and detect a light-dark boundary in the window as a hole edge point. However, in the binarized screen, a disturbance image portion due to light reflected from the inner surface of the hole may appear as a bright portion image, so that the detection accuracy of the hole edge point deteriorates. On the other hand, as in the present embodiment, if the scanning line LSC, which is a measurement location, is set on the imaging screen with shading and the luminance distribution on the scanning line LSC is detected, the luminance change becomes gentle in the disturbance image portion. The disturbance image portion can be determined from the luminance change, that is, the differential curve, and the detection accuracy of the hole edge point Pb is improved.

The brightness in the bright and dark portions of the image screen is not uniform. If the brightness distribution curve on the scanning line LSC is differentiated over the entire area, a plurality of vertices appear in the differential curve. In some cases, it may be difficult to determine whether the hole corresponds to the edge of the hole. Therefore, in the present embodiment, a differential curve is created for a region where the luminance difference before and after the decrease is the largest among the regions where the luminance distribution curve continuously decreases, and the above processing is performed. This prevents erroneous detection due to the above. In this case as well, a plurality of vertices may appear due to noise or the like. However, since vertices due to noise are low, the detection accuracy can be ensured by excluding vertices below a predetermined threshold from processing targets.

When the position of the hole edge point Pb on each scanning line LSC is detected as described above, the x, y coordinates on the screen of each hole edge point Pb are calculated from the setting data of each scanning line LSC.
Next, a regression circle Cb (see FIG. 4D) that approximates the hole image b from the coordinates of the hole edge points Pb (regression processing is performed so that the total amount of deviation of each hole edge point Pb from the circle is minimized). Then, the center coordinates mx and my are obtained by using the center of the regression circle Cb as the center m of the hole image b. When obtaining the regression circle Cb, the shift amount of each hole edge point Pb with respect to the circle calculated by the regression processing is obtained, and when the shift amount of any of the hole edge points Pb is equal to or more than a predetermined value, the hole edge point Pb is determined. Is repeated and the regression processing is performed again until the deviation amounts of all the hole edge points Pb become equal to or less than a predetermined value.

The description of the method of detecting the center coordinates mx and my of the hole image b has been described above. Next, the method of detecting the x coordinate sx of the light-section image s will be described. First, a template TPs representing the fragment shape of the light-section image s is created from the teaching data by graphic processing, and a shaded imaging screen of the light-section image S shown in FIG.
As shown in (E), pattern matching is performed by the normalized correlation method or the like using the template TPs, and the approximate x coordinate sx ′ of the light section image s is determined. Note that, in order to reduce the processing time of the pattern matching, it is desirable to perform the pattern matching by reducing the imaging screen and the template TPs to the same ratio (for example, 1/4).

Next, as shown in FIG. 4 (F), based on the y-coordinate of the regression circle Cb at the determined x-coordinate sx ', the upper and lower sides of the regression circle Cb are parallel to the x-axis, respectively. A plurality (for example, three) of scanning lines LSC, which are important measurement locations, are set at a predetermined pitch in the y-axis direction. According to this, each scanning line L
The SC surely crosses a predetermined portion of the light section image s. Next, from the luminance distribution on each scanning line LSC, the coordinates of an image point Ps that matches each scanning line LSC of the light-section image s are detected,
From the coordinates of these image points Ps, as shown in FIG.
An equation of a regression line Ls (a line obtained by regression processing so as to minimize the total amount of deviation from the line of each image point Ps) approximating the light section image s is calculated, and the x coordinate of the regression line Ls is calculated. The x-coordinate sx of the light-section image s is set.
If the work A is normal, the image line Ls is parallel to the y-axis. However, if the work A is inclined with respect to the XY coordinate plane of the space coordinate system, the image line Ls is inclined with respect to the y-axis. . In this case, an abnormal display is performed, and the x coordinate of the intersection of the image line Ls and a straight line parallel to the x axis passing through the center of the regression circle Cb is obtained as a temporary x coordinate sx of the light section image s.

It is to be noted that a binarized image is formed by binarizing the imaging screen with shading, and a long window is set in the x-axis direction as a measurement point at the position of the scanning line LSC, and the image center of gravity in the window is set. Can be determined as the image point Ps, but the image portion due to noise or disturbance light may appear as a bright portion together with the light cut image on the binarized screen, so that the detection accuracy deteriorates.

Therefore, in the present embodiment, as shown in FIG. 6A, a luminance distribution graph representing the luminance distribution on the scanning line is created, and the peak degree P of the luminance change at each point on the scanning line is obtained from this graph. The coordinates of the image point Ps that matches the light section image s are detected based on the peak degree P. The peak degree P of a point on the scanning line whose x coordinate is n is defined as an an luminance point on the luminance distribution graph at the point, and luminance at both ends of a measurement range W of a predetermined width set on the scanning line with the point as the center. The luminance points on the distribution graph are defined as bn and cn, respectively, and are obtained as values representing the height of an with respect to the connection line connecting bn and cn. In this case, the height of an with respect to the middle point of the connection (= {2an− (bn + cn)} / 2) may be used as the peak degree P, and the length of a perpendicular line descending from an in the connection is defined as the peak degree P. It is good. Here, FIG.
, The peak degree P at the point X = n1 has a positive value, and the peak degree P at the point X = n2 has a negative value. Then, a zero point P0 at which the peak degree becomes zero appears on both sides of the peak of the peak degree distribution curve, and a section between the zero points P0 is a part which is not affected by the background illumination by the illuminator on the line. Therefore, as shown in FIG.
If the area centroid G ′ in the section between 0 is obtained and the position is set as the position of the image point, the position of the image point can be accurately detected without the influence of the background illumination.

However, the inclination of the peak of the brightness distribution graph on the imaging direction side (right side) tends to be gentler than the inclination of the opposite side, so that the area center of gravity G 'of the peak is the position of the peak of the peak. From the camera to the imaging direction. On the other hand, the peaks of the peak degree distribution graph have substantially the same inclination on both sides even if the inclinations of the peaks of the luminance distribution graph are different on both sides. Therefore,
The area center of gravity G of the peak in the section between the zero points P0 of the peak degree distribution graph does not shift to the left from the position of the peak of the peak.
Therefore, in order to improve the detection accuracy of the image points, the area centroid G of the section between the zero points P0 of the peak degree distribution graph is obtained,
It is desirable that the position of the center of gravity G be the x coordinate of the image point Ps.

Incidentally, depending on the measurement site of the work A, as shown in FIG.
P exists or the work A as shown in FIG.
Hole B is provided in the stepped seating surface AS of FIG.
As shown in 1), the hole B may be formed of a cylindrical member AC such as a collar or a nut welded to the work A. At the measurement site in FIG. 7 (A-1), the light-section image s also appears inside the hole as shown in FIG. 7 (A-2), and in the measurement site in FIG. 7 (B-1). 7 (B-2), the light-section image s has a stepped shape, and at the measurement site in FIG. 7 (C-1), the light-section image s as shown in FIG. 7 (C-2). Is divided into an image corresponding to the surface of the work A and an image corresponding to the end surface of the tubular member AC.

When pattern matching using the template TPs is performed on the imaged screen on which the light-section image s appears, the template TPs matches the image in the hole in FIG. In FIG. 7 (B-2), the template TPs matches the image corresponding to the work general surface outside the seating surface, and in FIG. 7 (C-2), the template TPs matches the image corresponding to the end surface of the tubular member AC. In some cases, it is not possible to determine whether the template TPs matches the light-section image s present in the predetermined measurement range of the work A, and the displacement dZ of the work A cannot be accurately detected.

Therefore, in the present embodiment, the attribute information indicating the measurement range in which the work surface whose position is to be measured exists with reference to the center of the hole B, that is, the data of the inner diameter RIN and the outer diameter ROUT of the measurement range is taught. Measured and stored at the time, and when measuring at each measurement site, FIG. 7 (A-3)
(B-3) As shown in (C-3), the inside of the circle having a radius of RIN and the outside of the circle having a radius of ROUT centered on the hole center m obtained as described above from the hole image b are masked. Then, pattern matching is performed in this state.
According to this, the template TPs matches the portion of the light section image s corresponding to the predetermined measurement range. Then, by setting the scanning line LSC as described above between the y coordinate of the circle having the radius of RIN and the y coordinate of the circle having the radius of ROUT in the approximate x coordinate of the light-section image s obtained from the pattern matching. Thus, the x coordinate sx of the light-section image s corresponding to the work surface in the predetermined measurement range can be correctly obtained, and the displacement dZ of the work A can be accurately detected. Note that, at the measurement site shown in FIG.
7 is set based on the diameter of ROUT, and there is no ROUT.
At the measurement site (B-1), RIN and ROUT are
7C is set based on the inner and outer diameters of the S. At the measurement site in FIG.
Set based on the diameter of the mounting hole.

[Brief description of the drawings]

FIG. 1 is a perspective view showing an outline of an optical measuring device.

2A is a diagram viewed from the Y-axis direction in FIG. 1, and FIG. 2B is a diagram viewed from the X-axis direction in FIG.

FIG. 3A is a diagram illustrating an imaging screen of a hole image, and FIG. 3B is a diagram illustrating an imaging screen of a light-section image.

FIG. 4A is a diagram showing a differentiation screen of a hole image, and FIG.
A diagram showing a situation of pattern matching, (C) a diagram showing setting of scanning lines, (D) a diagram showing a regression circle approximating a hole image,
(E) A diagram showing the state of pattern matching of the light-section image with respect to the imaging screen, (F) a diagram showing the setting of scanning lines,
(G) Diagram showing image lines

5A is a diagram illustrating a luminance distribution on a scanning line in FIG. 4C, and FIG. 5B is a diagram illustrating a differential curve and an approximate parabola of the luminance distribution.

6A is a diagram illustrating a luminance distribution on a scanning line in FIG. 6B, and FIG. 6B is a diagram illustrating a method for detecting image points;

FIG. 7 (A-1), (B-1), and (C-1) are diagrams showing the cross-sectional shapes of respective measurement sites, (A-2), (B-2), and (C-2).
The figure which shows the imaging screen of the light-section image of each measurement part, (A-
3) (B-3) (C-3) A diagram showing a masked imaging screen

[Explanation of symbols]

2 Imager 4 Image processing circuit A Work B Hole b Hole image TP template m 'Approximate center LSC Scan line Pb Hole edge

Claims (3)

[Claims] 1. A method for optically measuring a position of a hole provided in a work, comprising: a first step of calculating a rough center position of a hole image appearing on an image pickup screen when the work is imaged; A second step of setting a plurality of measurement points so as to intersect a plurality of points of the hole edge of the hole image with reference to the obtained approximate center position, and a hole edge matching each measurement point of the hole edge of the hole image And a third step of detecting the coordinates of a point, wherein the second step is configured to set the plurality of measurement points along a plurality of radiation directions centered on the approximate center position, An optical measurement method of a hole position, characterized in that: 2. The method according to claim 1, wherein each of the measurement points is in each of the radiation directions.
The scanning line is set as a scanning line extending over a predetermined range determined based on a hole diameter stored in advance, and in the third step, each of the hole edge points is determined based on a luminance distribution on each of the scanning lines of an imaging screen with shading. The optical position measuring method according to claim 1, wherein the coordinates of the hole are detected. 3. The first step is to create a differentiated screen obtained by differentiating the luminance distribution of an imaged screen with shading, and to generate a differentiated screen for the differentiated screen using a template representing the shape of the hole edge of the hole image. By performing pattern matching,
3. The method of optically measuring a hole position according to claim 1, wherein the method is configured to determine an approximate center position of the hole image.