JP2523420B2 - Image processing method in optical measuring device - Google Patents

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an image processing method for an optical measuring apparatus provided with a projector for irradiating a work with slit light and an imager for picking up a light section image drawn by the slit light irradiated on the work. .

[0002]

2. Description of the Related Art Conventionally, an optical measuring device of this kind is used,
It is known to measure the shape and position of a work from a light section image on the screen of an imager. In this case, with the horizontal and vertical coordinate axes on the screen as the X-axis and the Y-axis, respectively, the light-section image has a maximum portion in one coordinate axis, for example, the X-axis direction. If the point can be clearly identified, the shape and position of the workpiece can be measured with reference to this maximum point, but if the maximum point is rounded, it will be difficult to uniquely determine the maximum point. , Calculate the equations of the lines of the light-section image located on both sides in the Y-axis direction with respect to the maximum part, and determine the position of the intersection of the line of one side of the image in the Y-axis direction and the line of the other side of the image. It is considered that the workpiece is measured by using this equation as an alternative to the maximum point.

[0003]

By the way, the light section image becomes a band-shaped image having a certain width, and when the equations of the lines of the image portion on one side and the other side in the Y-axis direction are calculated as described above, A plurality of windows are set in this image portion, the barycentric position of the light-section image in each of these windows is measured, and an equation of a curve or a straight line passing through these barycentric positions is calculated. When the position of the workpiece changes, the position of the light-cutting image on the screen also changes, so if the window is set to a fixed position on the screen, the window may deviate from the light-cutting image or the light-cutting may occur on individual workpieces. Since the set position of the window with respect to the image is displaced, the set position of the window needs to be displaced in accordance with the displacement of the light section image. In this case, it is conceivable to detect the displacement of the image from the position of the maximum portion of the light-section image and change the setting position of the window. However, if the maximum portion is rounded as described above, the X-axis of the maximum portion is detected. Even if the displacement in the direction can be detected to some extent accurately, the displacement in the Y-axis direction cannot be detected accurately, so that the set position of the window may shift in the Y-axis direction. The present invention has been made in view of the above points, and an object thereof is to provide an image processing method capable of accurately setting a window at a required portion of an image even if the light-section image is displaced on the screen.

[0004]

In order to achieve the above object,
The present invention is an image processing method for an optical measuring device, which comprises a light projector that irradiates a work with slit light, and an imager that captures a light-section image drawn by the slit light with which the work is irradiated. The upper light-section image has a maximum in the direction of one coordinate axis of the screen, and a predetermined window is set in each of the parts of the light-section image located on both sides of the other coordinate axis of the screen with respect to the maximum, In measuring the barycentric position of the light section image in each window, the position of the extreme end of the maximum section of the light section image in the direction of the one coordinate axis is measured, and a predetermined length is returned from the position in the direction of the one coordinate axis. At the position, two windows that are long in the other coordinate axis direction are set, the barycentric positions of the light-section images in each of the two windows are measured, and a reference point having a predetermined correlation with the barycentric positions is set. Ask, said And setting a constant of the window with respect to each said reference point in a predetermined positional relationship.

[0005]

The operation of the present invention will be described with the above-mentioned one coordinate axis as the X axis and the other coordinate axis as the Y axis. When the maximum portion of the light-section image is rounded, a plurality of pixels in the Y-axis direction detect the extreme position of the maximum portion, and the extreme Y-axis coordinate value cannot be uniquely determined. The axis coordinate values are uniquely determined,
If two windows, which are long in the Y-axis direction, are set at positions where the X-axis coordinate values are moved back by a predetermined length in the X-axis direction, the two windows are moved backward in the X-axis direction from the maximum portion and one side and the other side in the Y-axis direction are set. Each window hangs on a predetermined portion of the light-section image that extends to. Then, the image barycentric position in both windows is measured, and if a point having a predetermined correlation with both barycentric positions, for example, the midpoint of both barycentric positions is used as the reference point, the X and Y coordinate values of this reference point are The displacement of the light section image in the X-axis direction and the Y-axis direction on the screen is almost accurately represented. Therefore, even if the light-section image is displaced in the X-axis and Y-axis directions on the screen, the predetermined points are set in the Y-axis direction one side portion and the other-side portion with respect to the maximum portion of the light-section image based on the reference points. The window of can be accurately set with respect to these image parts with a fixed positional relationship. Therefore, it becomes possible to accurately calculate the equation of the image line of the one side portion in the Y-axis direction and the equation of the image line of the other side portion from the image centroid position measured in these windows, and to obtain both image lines obtained from the equations. The shape and position of the workpiece can be measured with high accuracy based on the intersection point of.

[0006]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS The illustrated embodiment is an optical measuring device for measuring the position of the work A based on a light section image of the work A formed of the roof side rails of the vehicle body shown in FIG. The present invention is applied to the image processing of
As shown in FIG. 2, the measuring device includes a projector 1 including a laser or the like for irradiating a work A with vertical slit light,
The image pickup device 2 formed of a CCD camera for picking up the light section image S ′ of the work A, which is an image of the slit light on the surface of the work A, and the optical axis of the slit light and the optical axis of the image pickup device 2 have a required angle θ. The base plate 3 is attached to the common base plate 3 in such a positional relationship as to be obliquely crossed. Reference numeral 4 in the figure is an electronic control circuit for performing image processing.

As shown in FIG. 3, the screen W of the image pickup device 2 has a maximum part in the X-axis direction with the horizontal and vertical coordinate axes of the screen W as the X-axis and the Y-axis, respectively. Y for department
A light-section image S having a curved portion on the upper side in the axial direction and a linear portion on the lower side is formed. In this image S, since the maximum portion is rounded and there is no clear corner portion suitable for the position determination reference,
Screen W of the intersection Q of the curved line S ₁ and the straight line S ₂
The above X and Y coordinate values are calculated from the equation of the curve S _{1 and} the equation of the straight line S ₂ , and the position of the work A is measured with the intersection Q as a reference point.

Here, the equation of the curve S ₁ can be approximately expressed as an equation of a circle passing through the three points in the specific region of the curved portion, and therefore the first to third portions of the upper half of the light section image S are represented. 3 windows W1, W2, W3 are set, and the centers of gravity G1, G of the light section images S in these windows are set.
2. The X and Y coordinate values of G3 were measured, the equation of a circle passing through these three centers of gravity was determined, and this was used as the equation of the curve S ₁ . By the way, the light section image S has a certain width, and the coordinate value of the center of gravity changes due to a subtle difference in whether or not the amount of light received by the pixels located near the boundary line of the image S exceeds a predetermined threshold value. Even if the same work is placed at the same position, the coordinate value of the center of gravity changes slightly. In this case, comparing the change rate of the radius of the circle with respect to the change of the coordinate value of the image center of gravity G2 of the intermediate second window W2 and the change rate of the radius of the circle with respect to the change of the coordinate values of the other center of gravity G1 and G3, The former rate of change is larger, and the error in the coordinate value of the center of gravity G2 greatly affects the calculation accuracy of the equation of the curve S ₁ . On the other hand, the curvature of the work A does not fluctuate to the left for individual products. If this curvature radius is measured in advance, the equation of the curve S ₁ can be calculated from the coordinate values of the two points on the curve S ₁ and the curvature radius. Can be calculated. When the curve equation is sampled and calculated from a minute area, in order to eliminate the influence of the error of the coordinate value of G2 and improve the equation calculation accuracy, the curve S ₁ is calculated from two points and the radius of curvature. It is advantageous to calculate the equation,
Therefore, in this embodiment, as shown in FIG.
The first and third windows W1 and W3 are set in the upper half part of each of the windows, the coordinate values of the image centroids G1 and G3 of the respective windows are measured, and from this and the radius of curvature R, the equation of the curve S ₁ is obtained. Was calculated. When measuring the radius of curvature R,
A master work having the same shape as the work A is placed at a predetermined reference position, a light-section image of the master work is taken, and the three first to third windows W1, W2, and W3 are set on the screen W, and these windows are set. The radius of the curve of the light-cut image of the masterwork is obtained from the image center of gravity, and the average value of the radii obtained by repeating this several times is set as the radius of curvature R, and this value is stored in the electronic control circuit 4. In the case of performing online measurement on a production line, the master work may be a vehicle body flowing through the line. In that case, the first one or several machines at the time of starting the measurement are master works. The equation of the straight line S ₂ of the light-cut image S of the work A is set with the fourth and fifth windows W4 and W5 in the lower half of the image S, and the coordinate values of the image centroids G4 and G5 of the respective windows are set. It is calculated by measuring.

By the way, when the relative positional relationship between the work A and the image pickup device 2 changes, the position of the light section image S on the screen W also changes, and the windows W1 to W5 may be set to the fixed positions of the screen W. , The window is out of the light section image S,
Further, the windows W1 and W3 set in the curved portion may be applied to the image portion having a small radius of curvature, and the curve equation may not be accurately calculated. In this case, as shown in FIG. 4, it is conceivable that the positions of the windows W1 to W5 are set with reference to the leftmost end point P of the maximum portion in the X-axis direction of the light section image S, but the light section image S It is difficult to unambiguously determine the Y-axis coordinate value of the point P unless the maximum part of the is not squared, and the relative positions of the windows W1 to W5 with respect to the light section image S vary in the Y-axis direction. . In view of this, in the present embodiment, as shown in FIG. 5A, the sixth and seventh upper and lower parts that are long in the Y-axis direction are provided at positions separated by a predetermined length to the right with respect to the X coordinate value of the point P. Set a pair of windows W6 and W7,
The coordinate values of the image centroids G6 and G7 of the respective windows are measured, and the coordinate value of a point having a predetermined correlation with the centroids G6 and G7, for example, the coordinate value of the reference point M which is the midpoint of the centroids G6 and G7, is calculated. Based on this point M, the positions of the windows W1 to W5 are determined as shown in FIG. 5 (b).

According to this, even if the position of the light cut image S on the screen W is displaced, the windows W1 to W5 can be accurately set while maintaining a constant correlation with the image S, and thus the curve S ₁ And the straight line S ₂ are correctly calculated, and the coordinate value of the intersection Q can be accurately obtained. In FIG. 5B, dxn and dyn are the positions in the X-axis direction and the Y-axis direction with respect to the reference point M of the upper left corner of each window Wn, Wxn,
Wyn is the length of each window Wn in the X-axis direction and the Y-axis direction, and is stored in the electronic control circuit 4.

Further, when the work A is displaced in the optical axis direction of the slit light, the light section image S is displaced in the X axis direction on the screen W and the enlargement ratio of the image S is changed. This will be described with reference to FIG. In FIG. 6, the optical axis of the image pickup device 2 is the Z axis,
The image pickup device 2 is placed at a point C on the Z axis, the X axis and the Y axis are taken on the reference image pickup surface W ′ of the image pickup device 2 which is separated from the point C by a predetermined reference distance L, and the light of slit light is emitted. A U-axis in the optical axis direction and a V-axis orthogonal to the optical-axis are arranged on a light-cutting plane T parallel to the slit light including the axis so that the V-axis coincides with the Y-axis and XY
It shows a state in which the coordinate system is set so that the U axis passes through the origin O of the coordinates. Now, let us say that the image length when an image of length H on the light section plane T, which is located at a distance a in the Z-axis direction from the image plane W ′, is projected onto the image plane W ′ with the point C as the viewpoint. h
Then, the enlargement ratio K = h / H is L / (La).
Here, assuming that the U-axis coordinate value of H is u and the X-axis coordinate value of h is x (the left side of the origin O is positive and the right side is negative), a is ucos θ.
Therefore, the enlargement ratio K is K = L / (L-ucos θ) (1) On the other hand, tan α = x / L, where α is the angle between the Z-axis and the line of sight from point C to H on the X-Z coordinate plane, and the X-axis coordinate value of H is usinθ, so tan α = usinθ / (L−a) = usinθ / (L−ucosθ), from which the relationship between x and u is calculated: x / L = usinθ / (L−ucosθ) That is, u = Lx / (Lsinθ + xcosθ). 2), and by substituting the equation (2) into the equation (1) and rearranging, the enlargement factor K is K = 1 + (x / L) cotθ (3). Therefore, when the work A is displaced in the optical axis direction of the slit light, the light section image S is displaced in the X axis direction on the screen W, and the displacement direction is left when the work A approaches, and right when the work A moves away. Moreover, the size of the image S is enlarged or reduced according to the enlargement ratio of the equation (3).
Therefore, the window W1 to
It is desired to change the position and size of W5. Therefore, in the present embodiment, when the work A is set to the normal position, the optical measuring device is adjusted so that the X-axis coordinate value on the screen W of the point P at the leftmost end of the light section image S becomes zero. Incidentally, when the image S is displaced in the X-axis direction, the X-axis coordinate value Px of the point P is substituted as x in the equation (3) to obtain the enlargement factor K,
As shown in FIG. 7, the Y-axis coordinate value of each window Wn is set to the above d.
The value of yn is multiplied by K, and the size of each window Wn is changed according to the enlargement ratio K so that the windows W1 to W5 can be set at the required positions of the light section image S, respectively.

Further, in the measurement of the shape by image processing, an error due to the distortion of the lens system of the image pickup device 2, an error due to focus shift, a processing error of the base plate 3 and an attachment error of the projector 1 and the image pickup device 2 to the error. An error may occur due to an error in the relative positional relationship between the projector 1 and the image pickup device 2 caused by the error. Therefore, as shown in FIG.
A table 6 which is mounted on the base 5 and is movable in two orthogonal directions on the light cutting plane, that is, the U-axis direction and the V-axis direction, is provided on the surface plate 5, and the table 6 resembles the shape of the work A. The reference block 7 having a shape was placed, and the error was measured in advance. To explain this in detail, the reference block 7 is set to the table 6
Therefore, the virtual corner points of the reference block 7 corresponding to the intersections to be described later are set to a plurality of measurement points U ₁ set on the light cutting plane.
V ₁ , ... Umvm are sequentially moved so as to match each other, and at each point, the reference block 7 is irradiated with slit light from the light projector 1 and a light-section image S ′ on the block 7 is picked up by the image pickup device 2 and picked up. From the light section image on the screen of the container 2, the position of the intersection (the point corresponding to the above Q) of the two lines forming the image is obtained by the same method as above, and the coordinates on the light section plane corresponding to this intersection point are obtained. The errors Δu and Δv in the U-axis direction and the V-axis direction between the value and the coordinate value of the measurement point are measured for each measurement point.
In this embodiment, an edge portion 7a that coincides with the virtual corner point is formed at the base of the reference block 7, and the reference block 7 is accurately measured by using the edge portion 7a as a guide so that the virtual corner point coincides with the measurement point. I was able to set to. FIG. 9 is a diagram three-dimensionally showing each Δu at a plurality of measurement points U ₁ V ₁ , ... UmVm, and a correction formula Δu = f representing Δu by multiple regression from the data of Δu.
₁ (u, v) is calculated, and a correction expression Δy = f ₂ that similarly expresses Δv
(U, v) is obtained, and the correction value obtained by these correction equations is added to or subtracted from the coordinate value on the optical cutting plane of the intersection Q obtained by the measurement of the work A to obtain the coordinate value for determining the position of the work A. ,
The errors due to the above various causes are collectively corrected so that accurate measurement can be performed.

A summary of the above processing procedure is shown in FIG.
It becomes the street. That is, first, the correction formula f ₁ ,
f ₂ is obtained, then the radius of curvature R is obtained using the master work, and then the mass production vehicle body is measured. In this measurement, as shown in FIG. 11, first, the position of the leftmost point P of the light section image S is measured (S1), and then the sixth point is set with reference to the point P.
The window W6 is set (S2), and the position of the image center of gravity G6 on the window W6 is measured (S3). Further, the seventh window W7 is set based on the point P (S4), and the position of the image center of gravity G7 on the window W7 is measured (S4).
5) Then, the reference point M is determined from the positions of the center of gravity G6 and G7 (S6). Next, the first window W1 is set on the basis of the reference point M, the position of the image center of gravity G1 of the window W1 is measured (S7, S8), and the third to fifth windows W3, W4, W5 are similarly processed. Are sequentially set and the positions of the image centroids G3, G4, and G5 of each window are measured (S9 ... S14).
Here, the setting processing procedure of each window W1 to W5 is as shown in FIG. 12, that is, the parameters dxn, dyn, Wxn, indicating the position and size of each window Wn from the memory.
Wyn is read (), then the enlargement factor K corresponding to the X-axis coordinate value Px of the point P is obtained (), and dyn is K · dyn, W
xn and Wyn are changed to K · Wxn and K · Wyn, respectively (), and Wxn, dxn, dyn at the position from the reference point M to dxn, dyn
A window Wn having a size of Wyn is set (). FIG.
After returning to 1 and measuring the positions of the respective centers of gravity G1 to G5,
The equation of the curve S ₁ is calculated from 1 and G3 and the radius obtained by multiplying the radius of curvature R by the enlargement factor K (S15), and then G
Calculate the equation of the straight line S ₂ from 4 and G5 (S16),
The position of the intersection Q is obtained from both equations (S17), then the coordinate value of the screen W of the intersection Q is converted into the coordinate value on the light cutting plane (S18), and this coordinate value is calculated from the correction formulas f ₁ and f _2. The calculated correction values are added and subtracted (S19), and the coordinate values thus corrected are transmitted as position data of the work A to a host computer that determines the assembly accuracy of the vehicle body (S20), and one measurement process is completed. To do.

[0014]

As is apparent from the above description, according to the present invention, even if the light section image is displaced, the windows can be accurately set at the predetermined portions of the image portions on both sides of the enlarged portion, and the images on both sides can be set. There is an effect that the equation of the image line of the part is accurately calculated and the work can be measured with high accuracy based on the intersection of both image lines.

[Brief description of drawings]

FIG. 1 is a perspective view of a work.

FIG. 2 is a plan view of an example of an optical measuring device to which the present invention is applied.

FIG. 3 is a diagram showing a method of obtaining an intersection of image lines on both sides of a maximum part of a light section image.

FIG. 4 is a diagram showing a method of obtaining an equation of an image line of a curved portion of a light section image by setting two windows.

5A is a diagram showing how to obtain a reference point according to the present invention, and FIG. 5B is a diagram showing setting parameters of a window based on the reference point.

FIG. 6 is a diagram showing the principle of image change due to a change in the optical axis direction of the work.

FIG. 7 is a diagram showing changes in window setting parameters according to image changes.

FIG. 8 is a perspective view showing a device layout when measuring an error.

FIG. 9 is a three-dimensional view of an error measured on a coordinate plane that matches a light-section plane.

FIG. 10 is a flowchart showing the overall measurement procedure.

FIG. 11 is a flowchart showing a processing procedure when measuring a workpiece.

FIG. 12 is a flowchart showing a window setting procedure.

[Explanation of symbols]

1 Light projector 2 Imager A Work W screen S Light section image W1 to W5 Windows W6 and W7 set on both sides of maximal part Window G1 to G7 set at a predetermined distance from the tip of maximal part Image center of gravity M Reference point Q Intersection point

Claims (2)

(57) [Claims] 1. An image processing method for an optical measuring device, comprising: a light projector for irradiating a workpiece with slit light; and an imager for capturing a light-section image drawn by the slit light irradiating the workpiece. The light cut image on the screen has a maximum in the direction of one coordinate axis of the screen, and a predetermined window is set in each of the parts of the light cut image located on both sides of the coordinate axis of the screen with respect to the maximum. In measuring the barycentric position of a light-section image in each window, the tip of the maximum position of the light-section image in the direction of the one coordinate axis is measured, and the position is returned from this position in the direction of the one coordinate axis by a predetermined length. At this position, two windows that are long in the direction of the other coordinate axis are set, the barycentric positions of the light-section images in each of the two windows are measured, and a reference point having a predetermined correlation with the barycentric positions. Ask for An image processing method in an optical measuring apparatus, characterized in that a predetermined window is set in a predetermined positional relationship with respect to the reference point. 2. A plurality of the predetermined windows are respectively set in portions of the light section image located on both sides in the other coordinate axis direction with respect to the maximum portion, and a plurality of the predetermined windows are set at one side of the light section image in the coordinate axis direction. While calculating the equation of the line of the one side portion from the position of the center of gravity of the light section image measured in the window of, the light section image of the light section image measured in a plurality of windows set in the coordinate axis direction other side portion of the light section image. The optical measurement according to claim 1, wherein the equation of the line of the other side portion is calculated from the position of the center of gravity, and the position of the intersection of the line of the one side portion and the line of the other side portion is obtained from both equations. Image processing method in device.