JPH054180A - Co-ordinate system matching method for multi-axis robot with hand-eye - Google Patents

Abstract

PURPOSE:To project the optical axis of the camera onto the robot co-ordinate system by determining the vector of the positional deviation between both the optical axes of the vision and the robot on the basis of a positional change in stationary reference point on the vision co-ordinate system due to rotation of the hand eye about the optical axis of the robot after parallelizing both the optical axes. CONSTITUTION:It is assumed that respective optical axes of the vision and the robot are adjusted to be parallel with each other. With a stationary reference point Ph being set as an origin Pro of the robot co-ordinate system, a hand eye 3 is moved using the point spaced by a specified distance L from the origin Pro along the optical axis of the robot as the camera position on the robot co-ordinate system, and then the hand eye 3 is moved orthogonally to the optical axis to thereby cause the optical axis of the vision to coincide with the stationary reference point Ph. The hand eye 3 is rotated by theta about the Y axis which is an orthogonally intersecting axis of the robot. Determination is made of a distance Z between the stationary reference point Ph previously calculated on the robot co-ordinate system and a stationary reference point Ph' recognized on the vision co-ordinate system. By Z which is the vector of the positional deviation, the position of the stationary reference point Ph on the robot coordinate system is shifted.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a method of aligning a visual coordinate system of a hand eye in a multi-axis robot with a hand eye and a robot coordinate system of the multi-axis robot itself. Of the camera position and the camera position on the robot coordinate system are eliminated.

[0002]

2. Description of the Related Art Conventionally, it has been proposed to mount a TV camera (hereinafter also referred to as a hand eye) having a built-in two-dimensional solid-state image pickup device on the hand of a multi-axis robot to support the handwork of the robot. In such a robot, a hand eye, that is, a visual coordinate system of a TV camera attached to the hand,
Work (hereinafter also referred to as calibration) to match the robot coordinate system of the multi-axis robot itself is required.

In the calibration of such a multi-axis robot with a hand eye, a visual coordinate system is defined as an orthogonal coordinate system having two orthogonal axes and an optical axis which are orthogonal to each other on an image pickup screen of the hand eye as coordinate axes. It is easy to align the origin of the visual coordinate system, that is, the camera position with the optical axis and the orthogonal axis on the robot coordinate system. Conventional caliber
In this section, the robot is taught the coordinates of a predetermined number of reference points as robot coordinate points on the robot coordinate system, and then these reference points are imaged from a predetermined position to obtain visual coordinate points on the visual coordinate system. , The coordinate relationship between robot coordinate points and their corresponding visual coordinate points (coordinate system transformation matrix)
Has been decided.

[0004]

The above-mentioned conventional calibration method can simply obtain the transformation matrix between the robot coordinate system and the visual coordinate system, but it requires a lot of manual work and the work is complicated. . Also, a camera optical axis on the visual coordinate system (hereinafter, also referred to as a visual optical axis or z axis),
The camera optical axis (hereinafter also referred to as the robot optical axis or Z axis) recognized by the robot coordinate system parallel to it is misaligned in a plane perpendicular to the robot optical axis, or the origin on the visual coordinate system, that is, the camera position. However, if there is a deviation in the robot optical axis direction with respect to the camera position recognized by the robot coordinate system in absolute space, a measurement error will occur.

The present invention has been made in view of the above problems, and eliminates the positional deviation between the actual camera position (that is, the origin position of the visual coordinate system) and the camera position on the robot coordinate system, thereby eliminating the calibrator. -The task to be solved is to simplify the operation (coordinate system matching) work.

[0006]

The first invention is a visual coordinate system having a visual optical axis of a hand eye fixed to a three-dimensionally movable hand as a coordinate axis and a position of the hand eye as an origin, and the hand. A robot coordinate system capable of calculating the position of the eye, and projecting the visual optical axis onto the robot coordinate system, that is, orthogonal to the visual optical axes of the robot optical axis as the calculation result and the visual optical axis. A coordinate system matching method for a multi-axis robot with a hand eye that eliminates positional deviation in the plane, in which the visual optical axis and the robot optical axis are adjusted in parallel, and while imaging a predetermined stationary reference point, A hand eye is rotated around the optical axis of the robot by a predetermined angle, and a positional shift vector between the optical axes on the visual coordinate system is based on a position change of the stationary reference point on the visual coordinate system due to the rotation. Determined, by displacing at least one of the position on the robot optical axis perpendicular direction of both the optical axis in accordance with the positional deviation vector is characterized in that to eliminate the positional displacement between the two optical axes.

According to the second aspect of the present invention, the visual coordinate system having the visual optical axis of the hand eye fixed to a three-dimensionally movable hand as the coordinate axis and the position of the hand eye as the origin, and the position of the hand eye can be calculated. A robot coordinate system, which is for projecting the visual optical axis onto the robot coordinate system, that is, for eliminating the positional deviation between the robot optical axis as the calculation result and the visual optical axis in the visual optical axis direction. A coordinate system matching method for a multi-axis robot with an eye, wherein the visual optical axis and the robot optical axis are adjusted in parallel, and the hand eye is centered on the stationary reference point while imaging a predetermined stationary reference point. The robot is rotated about an axis orthogonal to the optical axis of the robot by a predetermined angle, and based on the position change of the stationary reference point on the visual coordinate system due to the rotation, a positional shift between the optical axes on the visual coordinate system. Seek vector, by displacing at least one of the position of both the optical axis in accordance with the positional deviation vector in the direction of the robot optical axis, it is characterized in that to eliminate the positional displacement between the two optical axes.

According to the third aspect of the present invention, the visual coordinate system having the visual optical axis of the hand eye fixed to the three-dimensionally movable hand as the coordinate axis and the position of the hand eye as the origin, and the position of the hand eye can be calculated. Of a multi-axis robot with a hand eye that eliminates the positional deviation between the visual optical axis and the robot optical axis as a result of the projection of the visual optical axis onto the robot coordinate system. A coordinate system matching method, wherein a predetermined stationary gazing point at which the hand eye gazes in a predetermined posture from a position separated by a predetermined distance is used as a representative point on the robot coordinate system, and the visual coordinate system on the robot coordinate system is It is characterized in that the origin, that is, the camera position is defined as a displacement vector from the representative point.

[0009]

According to the above-described method of the first invention, based on the position change of the stationary reference point on the visual coordinate system by rotating the hand eye around the robot optical axis after parallelizing both optical axes. It calculates the displacement vector between both optical axes and shifts one of the optical axes so as to eliminate this displacement vector. By doing this, the actual optical axis of the camera can be easily placed on the robot coordinate system. And it is possible to project accurately.

In the method of the second aspect of the invention described above, after the two optical axes are made parallel, the stationary eye is penetrated, and the hand eye is rotated around an axis orthogonal to the robot optical axis, so that the optical axis between the two optical axes is increased. A position shift vector is obtained, and either optical axis is shifted so as to eliminate this position shift vector. In this way, the actual optical axis of the camera is projected onto the robot coordinate system easily and accurately. be able to.

In the above-described method of the third aspect of the invention, a predetermined gazing point at which the hand eye gazes in a predetermined posture from a position separated by a predetermined distance is set as a representative point on the robot coordinate system, and the visual coordinate system on the robot coordinate system is changed. The origin, that is, the camera position is defined as the displacement from the representative point. With this configuration, if the actual optical axis of the hand eye, that is, the visual optical axis is displaced from the position of the optical axis calculated by the robot coordinate system, that is, the robot optical axis, this positional displacement can be easily eliminated. An excellent effect can be achieved.

That is, in this case, the coordinates of the gazing point and the vector coordinates of the hand eye for picking up the gazing point are defined in advance on the robot coordinate system. On the other hand, from the image of the hand eye that gazes at this gazing point, the position and orientation of the hand eye with this gazing point as the origin can be known. Therefore, by comparing the vector coordinates of the hand eye obtained from the image of the hand eye (reference point is the reference point) with the vector coordinates of the hand eye preset on the robot coordinate system (reference point is the reference point), The positional deviation between them can be easily found, and further, the positional deviation can be easily eliminated by shifting the vector position of the hand eye on the robot coordinate system by the amount of the positional deviation between the two.

[0013]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENT A machine tool device using an embodiment of the present invention is shown in FIG. This device is a robot controller for control.
The multi-axis robot 1 is provided with a la (hereinafter, simply referred to as a controller) 10 and is capable of moving the hand 11 three-dimensionally, and a hand eye 3 attached to the hand 11 of the multi-axis robot 1. A lens system (not shown) having an optical axis 30 and a two-dimensional solid-state image pickup device (not shown) that is arranged orthogonal to the lens system to capture an image are stored.

The calibration plate 4 is arranged at a predetermined position just below the turning range of the hand 11. The calibration plate 4 is a thick metal plate having a hole 40, and the calibration described later, that is, the visual coordinate system of the hand eye 3 and the robot controller.
It is used for matching (correction) work with the coordinate system of robot (robot coordinate system). The hand eyes 3 are arranged directly above the calibration plate 4 with a predetermined distance therebetween.

Before carrying out the above calibration (coordinate system matching), the optical axis 30 of the hand eye 3 is aligned with the base axis A which is the center line of the hole 40, and the hand eye 3
And the optical axis and distance setting work for determining the distance between the hole 40 of the calibration 4 and the optical axis direction. The optical axis setting work is
This is because it is possible to eliminate the occurrence of measurement error due to inaccurate detection of the base axis A when the hand eye is moved in the calibration work performed later. Further, the distance setting work is to make the conversion between the pixel distance on the imaging screen and the actual distance on the surface of the hole 40 accurate.

Hereinafter, each setting operation will be described in order. (Direct Surface Out) First, the direct surface out will be described with reference to FIG. In FIG. 2, the calibration plate 4
Is disposed at a predetermined position on the flat floor surface, and a plane attachment 5 is disposed on the calibration plate 4.

The plane attachment 5 has a flat lower surface 53 and is mounted on the flat upper surface of the calibration plate 4, and a disc portion 55 projecting downward from the center of the disc portion 55. -Hole 40 of the option plate 4
And a lower shaft portion 54 that is fitted in close contact with the upper disc portion 55 and an upper shaft portion 56 that projects upward from the center of the disc portion 55. The upper shaft portion 56 and the lower shaft portion 54 are coaxial with the disc portion 55, and therefore, the attachment axis is the axis of the hole 40. .

The contours of the upper shaft portion 56 and the disc portion 55 are circles C1 and C2 when picked up from above, and the diameter of the circle C1 is designed to be d1 and the diameter of the circle C2 is designed to be d2 (see FIG. 3). A procedure for fully automatically performing this surface straightening using the robot controller 10 will be described below with reference to the flowchart of FIG. First, the hand 11 is controlled to move the hand eye 3 to a position substantially directly above the calibration plate 4, and the optical axis 30 is substantially aligned with the base axis A, so that the distance L between the attachment 5 and the hand eye 3 is adjusted. Reference distance L
It is made to approximately match o (100). In this embodiment, the optical axis 30 is located at the center pixel of the screen.

Next, the attachment 5 is imaged to capture an image signal (101), and a circle C is obtained from the obtained image signal.
1, C2 are extracted, they are normalized into a perfect circle to obtain normalized circles C1 ′ and C2 ′, and the center coordinate points (in the visual coordinate system) P1 (x1, of the normalized circles C1 ′ and C2 ′ are obtained. y1),
P2 (x2, y2) is calculated (102). It should be noted that these image processes are performed by an image processor (not shown) built in the robot controller 10.

Next, the hand 11 is moved so that the center point P1 of the normalized circle C1 'which is the inner circle is located at the center position (optical axis 30) of the screen (see FIG. 5) (104). Next, a straight line Lx (see FIG. 5) connecting the center coordinate points P1 (x1, y1) and P2 (x2, y2) is obtained, and the hand eye 3 is moved to a plane including this straight line Lx and the optical axis 30. Within the straight line Lx
(106). At this time, the swing angle of the hand eye 3 is set to an amount corresponding to the size of the straight line Lx (correctly, an amount having a trigonometric function relationship).

Next, the hand 11 is again moved so that the center point P1 is at the center position (optical axis 30) of the screen (see FIG. 5).
It is moved (108). Next, it is checked whether or not the length of the straight line Lx is smaller than the minute length ΔL. If it is smaller, it is determined that the surface straightening has been performed, and the routine is ended. If not so, the routine returns to step 106 to execute the routine. repeat.

With this configuration, since the attachment 5 suitable for the surface straightening is used, the surface straightening can be automatically executed with high accuracy. Further, in this embodiment, the optical axis 30 and the base axis of the hole 40 of the calibration plate 4 are aligned with each other, that is, the hole 40.
The optical axis 30 can be set at the center of the. Although the above-described embodiment describes the straight surface feeding using the straight surface attachment 50, the shape itself of the straight surface attachment 50 also has a great feature. That is, the attachment 50 has two surfaces that are parallel to each other and orthogonal to the axis, and concentric circles can be extracted from these two surfaces. Therefore, the surface straightening can be easily performed by the direction and length of the straight line connecting the center points of both circles.

Note that the attachment 5 can be locked to the calibration plate 4 not only by such a fitting structure but also by another method. For example, when the calibration plate 4 is a magnetic body, the attachment may be a magnet. FIG. 6 shows another aspect of the vertical attachment. In this attachment 5a, the upper shaft portion 5
6a has a frustoconical shape. In this case, it is preferable to illuminate the illumination light in a direction parallel to the optical axis 30.
In this way, the illumination light is totally reflected by the inclined surface 58a and the upper surface 51a. The 52a and the slope 58a can be visually distinguished. Of course, the slope 58a or the upper surface 51a. 52
Either of a may be colored. (Distance Out) Next, distance out will be described with reference to FIG.

After the above-described surface straightening and optical axis alignment, as shown in FIG. 8, the calibration plate 4 is
Place the distance attachment 6 on. The distance attachment 6 has a flat lower surface 63 and is mounted on the flat upper surface of the calibration plate 4, and a calibration plate projecting downward from the center of the disk portion 65. 4 has a lower shaft 64 fitted tightly into the hole 40, and a circular recess 66 is formed in the center of the disc portion 65 as an upper opening. Here, the recess 66 and the lower shaft 6
4, the disk portion 65 is coaxial, and therefore the axis of the attachment 6 is the axis of the hole 40, that is,
It matches the base axis A. The contour of the recess 66 becomes a circle C3 when the image is taken from above, and the diameter of the circle C3 is designed to be do (see FIG. 9).

This distance measurement is performed by the robot controller 10
The procedure shown in the flowchart of FIG.
With reference to FIG. First, image pickup is performed to capture an image signal (200), and the captured image signal is processed by an image processor to calculate an average diameter d of a circle C3 that is the contour of the recess 66 (202). Next, it is checked whether the difference (d-do) between the average diameter d and the true diameter do stored in advance is smaller than the minute value Δd (204), and if it is smaller, the routine is finished (204). Otherwise, d
Is checked to see if it is smaller than do (206). If it is smaller, the hand eye 3 is moved closer by a predetermined distance (210), and if not, the hand eye 3 is moved away by a predetermined distance (210).
Return to 204.

However, the true diameter do stored in the controller 10 is stored as a distance obtained on the screen when the hand eye 3 is positioned at the reference distance Lo, that is, a pixel distance, and the average. The diameter d is also displayed in pixel distance. As described above, in this embodiment, the distance from the attachment 6 of the hand eye 3 is detected by the size of the pixel distance of a straight line (diameter in this case) formed on the attachment 6 and having a known length. Therefore, it is possible to perform distance measurement with high accuracy and high speed.

The caliber, which is the main feature of this embodiment, will be described below.
Explain the work. This calibration work is carried out after the above optical axis setting and distance setting. (Optical Axis Matching) First, FIG. 11 to FIG.
14 and the flow chart of FIG.

First, as shown in FIG.
The optical axis 30 is aligned with the base axis A immediately above the motion plate 4, and the distance between the calibration plate 4 (here, its upper surface) and the hand eye 3 (here, its tip) is matched with the reference distance L. (400). The reference distance L is the distance set by the above-described distance setting, and the position of the hand eye 3 at this time is displayed as Pr1 on the robot coordinate system (see FIG. 12).

Next, the image of the calibration plate 4 is picked up to take in the image signal, and the hole 4 is obtained from the obtained image signal.
The center point of 0 (the stationary reference point in the present invention) is obtained, and the visual optical axis 30 is matched with this (402). It should be noted that in this embodiment, since the visual optical axis 30 is already aligned with the base axis A when the above-mentioned surface is directly projected, this step can be omitted.

Next, the coordinate point on the screen of the visual optical axis 30 is stored as the first visual coordinate point P1 (404, where
(See FIG. 13 for the first coordinate point P1). However, in this embodiment, since the visual optical axis 30 is already set at the center point of the screen by the above-described optical axis alignment, if the coordinate position is stored as the first coordinate point P1, this step can be omitted. it can.

Next, the hand eye 3 is moved along the optical axis of the robot by a predetermined distance (406). When the controller 10 supports the movement of the hand 11 by a predetermined vector on the robot coordinate system, the hand 11 is supposed to be displaced with respect to the vector. Hand eye 3 after moving
The position on the robot coordinate system is displayed as Pr2 (see FIG. 12).

Next, the calibration plate 4 is imaged again at this position Pr2, the obtained image signal is subjected to image processing, and the coordinate point of the center point of the hole 40 on the screen is set as the second visual coordinate point P2. It is stored (408), and then the hand eye 3 is returned to its original position (410). Next, the movement vector (see FIG. 13) between the first and second visual coordinate points P1 and P2 on the visual coordinate system is calculated (412). if,
This movement vector should be 0 if the robot optical axis and the visual optical axis match.

Next, it is checked whether the calculated movement vector is sufficiently small (whether it is smaller than a predetermined minute value ΔL (4
14) If it is smaller, it is assumed that the robot optical axis and the visual optical axis coincide with each other, and the process proceeds to the next routine (FIG. 17).
If the movement vector is not sufficiently small, the hand eye 3 is swung in the direction in which the movement vector decreases in size (416), and the process returns to step 406 again. The swing is performed in a plane including the robot optical axis and a vector in the robot coordinate system corresponding to the movement vector in the visual coordinate system, but may be performed in another plane. In this way, the visual optical axis can be easily and automatically matched with the robot optical axis by repeating the routine until the movement vector becomes sufficiently small.

(Orthogonal Axis Matching) Next, a predetermined visual orthogonal axis on the visual coordinate system which extends in a predetermined direction orthogonal to the visual optical axis (here, the X axis which is the horizontal scanning direction of the screen is selected).
And a robot orthogonal axis on the visual coordinate system that corresponds to the X axis and extends orthogonal to the robot optical axis (here, x
(Referred to as the axis) is shown in FIG.
5, FIG. 16 and FIG. 17 which is a flow chart.

Naturally, before starting the routine shown in FIG.
The robot optical axis, the visual optical axis 30, and the base axis A coincide with each other, the hand eye 3 is located at the position Pr1, and the first visual coordinate point P1 is stored. First, as shown in FIG. 15, the hand eye 3 is moved by a distance L in the x-axis (horizontal axis) direction which is one of the orthogonal axes of the robot (500), and the calibration plate 4 is imaged to obtain an image signal. The center point of the hole 40 (the stationary reference point in the present invention) is obtained from the acquired image signal, and the coordinate point on the screen is stored as the third visual coordinate point P3 (502, the third visual coordinate point). See Figure 16 for P3).

Next, the movement vector (see FIG. 16) between the first and third visual coordinate points P1 and P2 on the visual coordinate system is calculated (506). Next, the intersection angle Θ between this movement vector and the X axis as the visual orthogonal axis is obtained (508), and it is checked whether the intersection angle Θ is smaller than the minute value ΔΘ (510). If it is smaller, it is determined that the robot orthogonal axis x and the visual orthogonal axis X match each other, and the routine ends.

On the other hand, if the intersection angle Θ is equal to or greater than the minute value ΔΘ, the hand eye 3 is rotated in a direction in which the intersection angle Θ decreases in accordance with the size of the intersection angle Θ (512), and step 406 is performed again. Return. If the x axis of the robot orthogonal axis and the X axis of the visual orthogonal axis (horizontal scanning direction) match, the intersection angle Θ should be zero.

By doing so, the visual orthogonal axis and the robot orthogonal axis can be simply and automatically matched.
By the optical axis alignment work and the orthogonal axis alignment work described above, the both optical axes and the respective orthogonal axes substantially match (more precisely, become parallel). However, for example, as can be seen from FIG.
When the visual optical axis and the robot optical axis are parallel to each other and both are close to each other, even if the hand eye 3 is moved in the robot optical axis direction by a predetermined distance, the change in the movement vector is small. That is, with this method, it is not easy to accurately correct the positional deviation between the optical axes that are parallel to each other in the direction orthogonal to the optical axis.

Further, there is a possibility that the position of the hand eye, that is, the origin position of the visual coordinate system and that calculated by the robot coordinate system are misaligned in the optical axis direction, and this misalignment also needs to be corrected. A method of correcting the positional deviation between the origin of these visual coordinate systems and its position on the robot coordinate system (hereinafter also referred to as position matching), which is a feature of the present invention, will be described below. (Elimination of misregistration in the direction orthogonal to the optical axis) First, a method of eliminating misregistration in the direction orthogonal to the optical axis will be described with reference to the schematic diagram of FIG. 18, the stationary reference point movement diagram of FIG. 19, and the flowchart of FIG.

Before carrying out this step, it is assumed that all the above-mentioned steps have been carried out, the visual optical axis and the robot optical axis have been adjusted in parallel, and they are close to each other. First, in this embodiment, a stationary reference point Ph set at a predetermined position in absolute space is prepared. Note that, as the stationary reference point Ph, for example, the base axis A of FIG. 11 may be adopted. In this embodiment, the stationary reference point Ph is set to the origin P of the robot coordinate system.
As r0, the hand eye 3 is moved to the camera position on the robot coordinate system at a point distant from the origin Pr0 by a predetermined distance L in the robot optical axis direction (60).
0) Next, the hand eye 3 is moved in a direction orthogonal to the optical axis to match the visual optical axis (here, the visual optical axis is the center point of the image pickup screen) with the stationary reference point Ph (602). . The stationary reference point Ph is used as the origin of the robot coordinate system in order to simplify the calculation of the position of the camera position (the origin of the visual coordinate system) on the robot coordinate system. A point other than Ph may be used as the origin of the robot coordinate system.

Therefore, when the visual orthogonal axes are x and y axes, the visual optical axis is the z axis, the robot orthogonal axes are the X and Y axes, and the robot optical axis is the Z axis, the stationary reference point Ph is used.
The coordinates (x, y, z) on the visual coordinate system of are (0, 0, L
z) and the coordinates (X,
Y, Z) becomes (0, 0, 0). Lz is the distance in the visual optical axis direction between the stationary reference point Ph and the camera position in the visual coordinate system.

Then, the stationary reference point Ph = (0,0) on the xy plane on this visual coordinate system is stored in memory (6
04). Next, the hand eye is rotated 180 degrees about the robot optical axis (Z axis). (606). If the robot optical axis and the visual optical axis coincide with each other in the absolute space, the position of the stationary reference point Ph on the image pickup screen is not displaced. If they do not match, the stationary reference point Ph on the imaging screen is the stationary reference point P as shown in FIG.
It shifts to the position of h '(the position of (x1, y1) on the visual coordinate system) (608).

Therefore, the actual positions of the robot optical axis on the visual coordinate system are the points Ph (0,0) and Ph '(x1, y).
It becomes the position of the middle point (0.5x1, 0.5y1) with respect to 1) (610). V in FIG. 19 is the position shift vector in the present invention, and this position shift vector V is the stationary reference point Ph.
And the length of the line segment connecting Ph 'and Ph' to Ph '
With a direction towards.

Next, the camera position is re-designated on the robot coordinate system so that the camera position on the robot coordinate system, that is, the origin of the visual coordinate system is moved by the displacement vector V shown in FIG. 19 (612). Here, the original camera position is (0,0, L), so the new camera position is (-
0.5x1, -0.5y1, L). Where L is
Origin Pr0 of the robot coordinate system stored in the robot coordinate system
To the camera position.

Next, it is checked whether or not the absolute value (distance) of the displacement vector V is large, and if it is large, step 60 is executed.
Return to 2 and repeat the routine. Although the rotation angle is 180 degrees in the above embodiment, the rotation angle can be set freely. . (Elimination of misalignment in the optical axis direction) Next, elimination of misalignment between the actual position of the hand eye 3 in the optical axis direction (to be exact, the position on the visual coordinate system) and the calculated position on the robot coordinate system. A method for doing so will be described with reference to the schematic diagram of FIG. 21, the stationary reference point movement diagram of FIG. 22, and the flowchart of FIG.

Before carrying out this step, it is assumed that all of the above steps have been carried out and the visual optical axis and the robot optical axis have been adjusted in parallel. First, in this embodiment, the stationary reference point Ph is set as the origin Pro of the robot coordinate system, and a point separated from the origin Pr0 by a predetermined distance L in the robot optical axis direction is set as a camera position on the robot coordinate system.
Is moved (700), and then the hand eye 3 is moved in a direction orthogonal to the optical axis to match the visual optical axis (here, the visual optical axis is the center point of the imaging screen) with the stationary reference point Ph. (702).

The stationary reference point Ph is used as the origin of the robot coordinate system in order to simplify the calculation of the position of the camera position (origin of the visual coordinate system) on the robot coordinate system. A point other than the stationary reference point Ph may be the origin of the robot coordinate system. Therefore, now
When the visual orthogonal axes are x and y axes, the visual optical axis is the z axis, the robot orthogonal axes are the X and Y axes, and the robot optical axis is the Z axis, the coordinates of the stationary reference point Ph on the visual coordinate system. (X,
y, z) becomes (0, 0, Lz), and the coordinates (X, Y, Z) on the robot coordinate system become (0, 0, 0). Lz is the distance in the visual optical axis direction between the stationary reference point Ph and the camera position in the visual coordinate system.

Next, the hand eye 3 is rotated by Θ around the Y axis which is the robot orthogonal axis (704). Here, the Y axis passes through the stationary reference point Ph as the origin of the robot coordinate system as described above (Y in FIG. 21).
The axis passes through the stationary reference point Ph and stands at a right angle to the paper surface). In FIG. 21, if the actual distance (that is, the distance visually recognized on the visual coordinate system) Lz from the stationary reference point Ph to the camera position Pc0 before rotation is the camera position before rotation from the stationary reference point Ph recognized by the robot coordinate system. If the distance L to Pc0 is equal, the position of the stationary reference point Ph on the imaging screen should not be displaced by the above rotation. If they do not match, the stationary reference point Ph on the imaging screen
Shifts to the position of the stationary reference point Ph '(position (x2, y2) on the visual coordinate system) as shown in FIG.
This position is stored in memory (706).

Next, the distance ΔZ between the stationary reference point Ph calculated on the robot coordinate system and the stationary reference point Ph 'visually recognized on the visual coordinate system is obtained (708). As can be seen from FIG. 21, if the displacement distance of the stationary reference point Ph on the visual coordinate system (imaging screen) before and after rotation is ΔL, ΔZ is Δ
L / sin Θ. Next, the position of the stationary reference point Ph on the robot coordinate system is shifted on the robot optical axis by the above-mentioned ΔZ which is the position shift vector in the present invention. Naturally, the camera position, that is, the origin of the visual coordinate system is shifted by ΔZ in the visual optical axis direction on the robot coordinate system, and (0, 0, L
z-ΔL) (710).

Next, it is checked whether or not the distance ΔL is smaller than the predetermined distance. If the distance ΔL is smaller than the predetermined distance, it is determined that the positional deviation in the optical axis direction has been eliminated, and the routine ends. On the other hand, if it is not smaller, the routine returns to step 702 to execute the routine again. However, in this even-numbered routine, the position of the hand eye 3 is swung by rotating 2Θ to the right in step 704, while in the odd-numbered routine after the third time, it is moved to the left again in step 704. The position of the hand eye 3 is swung by rotating by 2Θ.

By doing so, the positional deviation in the optical axis direction can be corrected easily and accurately. Further, in this embodiment, the stationary reference point Ph described above is used as the robot representative point in this robot coordinate system, and the coordinates on the robot coordinate system are stored in advance. Further, the stationary reference point Ph is set as the gazing point of the hand eye 3. The point of interest here is
It means a point at which the hand eye 3 gazes in a predetermined direction from a camera position that is a predetermined vector distance away from the gazing point.

That is, in this embodiment, the hand eye 3
The camera position (the camera position or the origin of the visual coordinate system) is defined as the vector distance from the gazing point, and this gazing point is defined in the robot coordinate system as the robot representative point. It can be easily displayed as a position displaced by the above vector distance from the coordinates of the robot representative point.

In this way, when the camera position is displayed as a vector distance from the gazing point, the position of the gazing point on the robot coordinate system, that is, the robot representative point, is changed by the stationary reference point Ph on the imaging screen, that is, the displacement of the gazing point. By shifting, the camera position can be accurately corrected on the robot coordinate system.

[Brief description of drawings]

FIG. 1 is a schematic view of a multi-axis robot system using the setting method of the present invention,

FIG. 2 is a cross-sectional view showing a state in which a vertical attachment is attached,

3 is a plan view of the plane attachment of FIG. 2,

FIG. 4 is a flow chart showing a surface directing routine,

FIG. 5 is an explanatory diagram showing a screen,

FIG. 6 is a cross-sectional view showing a state in which another surface upright attachment is attached,

7 is a plan view of the surface attachment of FIG. 6,

FIG. 8 is a cross-sectional view showing a state in which a distance attachment is attached,

9 is a plan view of the distance attachment of FIG.

FIG. 10 is a flow chart showing a distance setting routine.

FIG. 11 is an explanatory diagram for explaining optical axis alignment,

FIG. 12 is an explanatory diagram for explaining optical axis alignment,

FIG. 13 is a screen diagram in optical axis alignment work,

FIG. 14 is a flowchart showing an optical axis matching routine,

FIG. 15 is an explanatory diagram for explaining orthogonal axis alignment,

FIG. 16 is a screen diagram in the orthogonal axis alignment work,

FIG. 17 is a flowchart showing an orthogonal alignment routine.

FIG. 18 is an explanatory diagram showing a positional deviation between both optical axes in a direction orthogonal to the optical axis;

FIG. 19 is a screen diagram showing the positional deviation of FIG.

FIG. 20 is a flow chart showing the position deviation eliminating routine of FIG.
Chart,

FIG. 21 is an explanatory diagram showing a positional shift in the optical axis direction between both optical axes;

FIG. 22 is a screen view showing the positional deviation of FIG. 21;

FIG. 23 is a flow chart showing the position deviation eliminating routine of FIG. 21.
Chart,

[Explanation of symbols]

3 is a hand eye, 4 is a calibration plate 4,
5 is a hand eye, 30 is an optical axis,

Claims (3)

[Claims] 1. A visual coordinate system having a visual optical axis of a hand eye fixed to a three-dimensionally movable hand as a coordinate axis and a position of the hand eye as an origin, and a robot coordinate system capable of calculating the position of the hand eye. A hand which has a projection of the visual optical axis on the robot coordinate system, that is, a positional deviation between the robot optical axis as the calculation result and the visual optical axis in a plane orthogonal to the visual optical axis. A coordinate system matching method for a multi-axis robot with an eye, wherein the visual optical axis and the robot optical axis are adjusted in parallel, and the hand eye is moved around the robot optical axis while imaging a predetermined stationary reference point. Rotate by a predetermined angle, determine the position shift vector between the optical axes on the visual coordinate system based on the position change of the stationary reference point on the visual coordinate system due to the rotation, to the position shift vector By displacing at least one position of both the optical axes in a direction perpendicular to the robot optical axis, the coordinate of the multi-axis robot with a hand eye characterized by eliminating the positional deviation between the both optical axes. System matching method. 2. A visual coordinate system having a visual optical axis of a hand eye fixed to a three-dimensionally movable hand as a coordinate axis and a position of the hand eye as an origin, and a robot coordinate system capable of calculating the position of the hand eye. And a projection of the visual optical axis onto the robot coordinate system, that is, a multi-axis with a hand eye that eliminates a positional deviation between the robot optical axis as the calculation result and the visual optical axis in the visual optical axis direction. A coordinate system matching method for a robot, wherein the visual optical axis and the robot optical axis are adjusted in parallel and the hand eye is imaged at a predetermined stationary reference point while the hand eye is centered on the stationary reference point. A predetermined angle is rotated around orthogonal axes, and a position shift vector between the two optical axes on the visual coordinate system is obtained based on a position change of the stationary reference point on the visual coordinate system due to the rotation. A multi-axis with a hand eye, wherein at least one position of the both optical axes is displaced in the direction of the robot optical axis according to the position displacement vector to eliminate the positional displacement between the both optical axes. Robot coordinate system matching method. 3. A visual coordinate system having a visual optical axis of a hand eye fixed to a three-dimensionally movable hand as a coordinate axis and a position of the hand eye as an origin, and a robot coordinate system capable of calculating the position of the hand eye. And a coordinate system matching method for a multi-axis robot with a hand eye that eliminates the projection of the visual optical axis onto the robot coordinate system, that is, the positional deviation between the robot optical axis as the calculation result and the visual optical axis. And
A predetermined stationary gazing point at which the hand eye gazes in a predetermined posture from a position separated by a predetermined distance is used as a representative point on the robot coordinate system, and the origin of the visual coordinate system on the robot coordinate system, that is, the camera position is the representative point. A coordinate system matching method for a multi-axis robot with a hand eye characterized in that it is defined as a displacement vector from.