JPH09196622A - Method for detecting position - Google Patents

Abstract

PROBLEM TO BE SOLVED: To markedly enhance correction accuracy insusceptible to robot controlling accuracy and resolution accuracy by a method wherein a distance between a visual sensing means and an object is changed and positions and angles of two points are measured three times so that correction method for the position and angle is formed. SOLUTION: In a position detecting method for an object by image processing, positions of two points of an object 1 and an angle of a line passing through the two points with respect to a reference line are obtained by image recognition using a visual sensing means 2. Next, the visual sensing means 2 is moved such that obtained positions and angle is equal to reference positions and a reference angle. On the moved position a, positions and angle of the two points are detected again. Next, a distance β between the visual sensing means 2 and object 1 are changed, then the positions of the two points and the angle are detected again, thereby correcting the positions and angle.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a method for detecting the position of an object by image processing, which is used when, for example, a vehicle door or other parts are assembled or removed by using a robot.

[0002]

2. Description of the Related Art Conventionally, as a method for detecting the position of an object to be measured by image processing (position detecting method), there is, for example, the method described in JP-A-2-243914. That is, as shown in FIG. 12, a master table 92 on which an object to be measured 91 is fixedly mounted, a CCD camera 93 as a visual means fixed in position with respect to the master table 92, and data captured by the CCD camera 93 are displayed. The master position data of the measurement point 96 with respect to the reference point 95 on the master table 92 is stored in the memory in the image processing device 94, and the measured object 91 is placed on the master table 92 in this state. Set to the specified position of
The positional relationship between the measurement point 96 of the DUT 91 by the CCD camera 93 and the reference point 95 of the master table 92 is calculated by the image processing device 94.

At this time, the master position data (X ₀ , Y ₀ ) of the measurement point 96 and the actually measured measurement point (X ₁ , Y 0
₁ ) and the respective reference points are made to coincide with each other by the calculation unit in the image processing device 94, and the errors ΔX = X ₁ −X ₀ and ΔY = Y ₁ −Y ₀ between these master reference position and actual measurement position are calculated,
This is a so-called coordinate position detection method in which the position coordinates of the measurement point are calculated to detect and correct the position.

However, in this conventional method, the correction is executed by reading only once, and the CCD camera 93 is used.
Since the angle information regarding is not read, the error is as large as about ± 2.5 mm, and there is a problem that it cannot be used for position detection that requires highly accurate correction of, for example, about ± 0.5 mm.

Generally, C as an image processing and visual means.
The resolution accuracy improves as the CD camera gets closer to the object to be measured, but when driving the robot in response to the position correction signal from the CCD camera, the angle component is considered in addition to the X-axis component and the Y-axis component. In addition to the fact that gravity acts on the robot side when the robot is driven, there are manufacturing errors of the robot, and there is a limit to the calculation capacity on the robot side, there is a movement deviation, so it is sufficient. However, there is a problem in that the high precision cannot be ensured and the robot operates at a position and angle different from the commanded position.

[0006]

The invention according to claim 1 of the present invention obtains the positions of two points of an object and the angle of a line connecting these two points with respect to a reference line by image recognition by visual means (1 After the next measurement), the visual means is moved so that the obtained position and angle become the reference position and the reference angle, and the positions and angles of the two points are re-detected (secondary measurement) at this moving position, Further, by changing the distance between the visual means and the object and re-detecting the positions and angles of the two points (third-order measurement), the method of correcting the position and the angle is achieved. An object of the present invention is to provide a position detection method capable of significantly improving correction accuracy without being influenced by robot control accuracy and resolution accuracy by detecting a total of three times.

According to a second aspect of the present invention, in addition to the object of the first aspect of the invention, the change in the distance between the visual means and the object is set so that the visual means approaches the object. An object of the present invention is to provide a position detecting method capable of improving the resolution (resolution accuracy) by this visual means.

According to the third aspect of the present invention, in addition to the object of the second aspect of the invention, the control system and the position detection function originally possessed by the robot side are effectively utilized by mounting the visual means on the robot. Then, it aims at providing the position detection method which can drive the above-mentioned visual means.

According to a fourth aspect of the present invention, in addition to the object of the third aspect of the invention, the position and angle are corrected by correcting the control error of the robot system. An object of the present invention is to provide a position detection method capable of controlling a robot arm at an appropriate position and angle with high accuracy.

[0010]

According to a first aspect of the present invention, there is provided a method for detecting the position of an object by image processing, wherein the position of two points of the object and the two points of the object are recognized by image recognition by visual means. The angle of the connecting line with respect to the reference line is obtained (primary measurement), and then the above-mentioned visual means is moved so that the obtained position and angle become the reference position and the reference angle. The angle and is detected again (secondary measurement),
Next, it is a position detecting method for correcting the position and the angle by detecting the position and the angle of the two points again (third measurement) by changing the distance of the visual means to the object.

According to a second aspect of the present invention, in combination with the configuration of the first aspect of the invention, the change in the distance between the visual means and the object is a position detecting method for bringing the visual means closer to the object. It is characterized by being.

According to a third aspect of the present invention, in addition to the configuration of the second aspect of the invention, there is provided a position detecting method in which the visual means is attached to a robot.

According to a fourth aspect of the present invention, in addition to the configuration of the third aspect of the invention, the correction of the position and angle is a position detecting method for correcting a control error of the robot system. And

[0014]

According to the invention described in claim 1 of the present invention, as shown in the claims correspondence diagrams in FIGS. 1 to 7, in the method of detecting the position of the object 1 by image processing, the image of the visual means 2 is detected. By recognition, the positions of the two points a and b of the object 1 (see the points on the image shown in FIG. 2) and the two points a and b
An angle θ of a line c connecting b (two points predetermined by the teaching point of the robot) with respect to a reference line d (specifically, a Y-axis reference line that the visual means has) is obtained (primary measurement), and then The above-mentioned visual means 2 is shown in FIG.
, The calculated position a = (x.y) and the angle θ
Is moved so as to become the reference position (0, 0) and the reference angle 0 degrees, and at this movement position (see the virtual line α in FIG. 1), two points e and f (points on the image shown in FIGS. 3 and 5 are shown. The position and the angle of (refer to) are detected again (secondary measurement).

At this time, when the primary measurement is correct, e = (0,0) and θ ₂ = 0 as shown in FIG. 3, and when the primary measurement is incorrect, e = (0,0) as shown in FIGS. 5 and 7. (X ₂ ,
y ₂ ), θ = θ ₂ (where θ ₂ ≠ 0). Next, the distance between the visual means 2 and the object 1 is changed to the approaching direction or the approaching direction (not shown) as indicated by the phantom line β in FIG. 1, and the positions and angles of the two points g and h are again determined. Detect (3rd measurement).

In this third-order measurement, when the above-mentioned first-order measurement is correct, g = (0,0) and θ ₃ = 0 as shown in FIG. 4, and when the first-order measurement is incorrect, FIGS. As shown in, g = (x ₃ , y ₃ ) and θ = θ ₃ (where θ ₃ ≠ 0). Furthermore, (x ₂ , y ₂ ), θ ₂ , (x ₃ , as position data and angle data of the secondary measurement and the tertiary measurement described above,
y ₃ ) and θ ₃ are used to correct the position and angle.

Thus, the above-mentioned primary measurement, secondary measurement, and 3
Since the correction (position on the X-axis, position on the Y-axis, axis deviation, angular deviation) based on the detection data is executed by detection of a total of three times in the next measurement, robot control accuracy and resolution accuracy There is an effect that the correction accuracy can be greatly improved without being affected by the above.

According to the invention described in claim 2 of the present invention,
In addition to the effect of the invention described in claim 1, since the change in the distance between the visual means and the object brings the visual means closer to the object, the resolution (resolution accuracy) of the visual means is improved. There is an effect that can be.

According to the third aspect of the present invention,
In addition to the effect of the invention described in claim 2, since the visual means is attached to the robot, there is an effect that the visual means described above can be driven by effectively utilizing the control system and the position detection function originally possessed by the robot side. .

According to the invention of claim 4 of the present invention,
In addition to the effect of the invention described in claim 3, the above-mentioned correction of the position and angle corrects the control error of the robot system. Therefore, by correcting the control error of the robot system, the robot arm can be adjusted to an appropriate position and angle. There is an effect that can be controlled with accuracy.

[0021]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS One embodiment of the present invention will be described below in detail with reference to the drawings. FIG. 8 shows a position detecting device used in the position detecting method of the present invention. The position detecting device 11 is a robot mount 12.
6-axis type robot 13 mounted on the robot hand, a robot hand 16 attached to the tip of the robot arm 14 of the robot 13 via a fastening portion 15, and a plurality of nut runners 17 provided on the robot hand 16. The CCD camera 2 as a visual means is attached to the nut runner 17 mounting portion via a bracket (not shown).
By the image recognition of the CD camera 2, the positions a and b of the two points of the object 1 (object to be measured) (see FIG. 2) and the angle θ of the line c connecting the two points a and b with respect to the reference line d (see FIG. 2). ) And is configured to ask.

A position detecting method using the position detecting device 11 thus constructed will be described in detail below. In the following description, a method of calculating a correction value after the above-described primary measurement, secondary measurement, and tertiary measurement (however, the following measurement is the same as the content already described in the section of the action and effect of the invention) Shown (replaced by robot calculation).

In FIG. 9, p1 indicates a command position during primary measurement, p2 indicates a command position during secondary measurement, p3 indicates a command position (target position) during tertiary measurement, and r1 indicates during primary measurement. , R2 indicates the take-in position during the secondary measurement, and r3 indicates the take-in position during the third measurement. Here, the components of the respective commanded positions p1, p2, p3 can be expressed by the following [Equation 1].

[0024]

[Equation 1]

In the above [Equation 1], x, y, and z represent the X axis, Y axis, and Z axis of the three-dimensional robot base coordinates,
"01", "02", and "03" are command positions p1, p2, p
3 indicates a command value, and tx, ty, and tz indicate angles around each axis, and ty particularly means the angle θ shown in the claims. Further, the components of the above-mentioned respective intake positions r1, r2, r3 can be expressed by the following [Equation 2].

[0026]

[Equation 2]

In the above [Equation 2], x, y, and z are the X-axis of the three-dimensional robot base coordinates, as in [Equation 1],
The Y-axis and the Z-axis are shown, and tx, ty, and tz also show the angles around each axis, and "1", "2", and "3" are the respective positions r1, r.
2 and r3 are actual values, and in particular ty means the angle θ shown in the claims. Each of the above command values p1, p2
The motion displacement vector of the robot 13 with respect to p3 is as shown in [Equation 3].

[0028]

[Figure 3]

The final motion correction value is the inverse vector of the above-mentioned shift, and should be the negative motion shift 3 originally, but even if the robot 13 is operated with the correction command value, gravity, manufacturing error, etc. will occur. The deviation occurs again due to the factor. Therefore, an arithmetic mean vector (median value of the two shifts 2 and 3) of the two obtained motion shift vectors (accurate motion shift 2 and motion shift 3) is obtained, and this value is approximately used as a correction value. To do. This correction vector can be expressed by the following [Equation 4].

[0030]

(Equation 4)

Here, r2 corresponding to the shift correction 2 in FIG.
When the vector of p2 is expressed by components, the following [Equation 5] is obtained.

[0032]

(Equation 5)

Further, r3p corresponding to the shift correction 3 in FIG.
When the vector of 3 is represented by components, the following [Equation 6] is obtained.

[0034]

(Equation 6)

In the above [Equation 5] and [Equation 6], “0” added before “p” and “r” indicates zero as the origin of the robot base coordinates. [Equation 5] and [Equation 6] are substituted into the above [Equation 4], and the correction vector is obtained by rearranging the vector formula.

[0036]

(Equation 7)

Therefore, the final command value (dest) to the robot 13 shown in FIG. 11 is a value obtained by adding the above correction vector to the vector of or3, and is represented by the following [Equation 8]. You can

[0038]

(Equation 8)

That is, when no correction is made, the coordinate can be moved in the direction from the coordinate origin 0 to r3, or by the above-described correction, it can be moved from the coordinate origin 0 in the direction p3 as the target position.

In summary, according to the position detecting method of the present invention, the positions of the two points of the object 1 by the image recognition of the visual means (see CCD camera 2) and the angle of the line connecting these two points with respect to the reference line (Equation 2) (See ty1 in the figure) (primary measurement), and the next visual means (see CCD camera 2) is moved so that the obtained position and angle become the reference position and the reference angle, and 2 at this moving position. The position and angle of the point (see ty2 in Formula 2) are detected again (secondary measurement).

Next, the distance between the above-mentioned visual means (see CCD camera 2) and the object 1 is changed, and the positions and angles (see ty3 in the equation 2) of the two points are detected again (third measurement). Further, the position and the angle are corrected (refer to the correction vector shown in Formula 7 and the vector of the final command value dest shown in Formula 8) based on the position data and the angle data of the above-mentioned secondary measurement and tertiary measurement.

As described above, in addition to the conventional method of only primary measurement, secondary measurement and tertiary measurement are additionally executed, and a total of three detections is performed to perform correction based on the detected data (position on the X axis, Since the correction of the position on the Y axis, the axis deviation, and the angular deviation) is performed, the correction accuracy can be greatly improved without being affected by the robot control accuracy and the resolution accuracy of the CCD camera 2. is there.

Further, since the above-mentioned visual means (see CCD camera 2) changes its distance from the object 1, the visual means is brought closer to the object 1, so that the resolution (resolution accuracy) of the visual means should be improved. There is an effect that can be.

Furthermore, the robot 13 has a visual means (CCD).
Since the camera 2 is attached, there is an effect that the above-mentioned visual means can be driven by effectively utilizing the control system and the position detection function originally possessed by the robot 13 side. in addition,
The above-described correction of the position and angle corrects the control error of the robot system. Therefore, by correcting the control error of the robot system, the bot arm 14 (including the nut runner 17 at its tip in the embodiment) is adjusted to an appropriate position and angle. There is an effect that can be controlled with accuracy.

In the correspondence between the structure of the present invention and the above-described embodiment, the visual means of the present invention corresponds to the CCD camera 2 of the embodiment, and the robot hereinafter corresponds to the 6-axis type robot 13. However, the present invention is not limited to the configurations of the above-described embodiments.

[Brief description of drawings]

FIG. 1 is a claim correspondence diagram showing a position detection method of the present invention.

FIG. 2 is a complaint correspondence diagram showing a primary measurement.

FIG. 3 is a complaint correspondence diagram showing a secondary measurement when the primary measurement is correct.

FIG. 4 is a complaint correspondence diagram showing a third measurement when the first measurement is correct.

FIG. 5 is a claim correspondence diagram showing a secondary measurement when the primary measurement is incorrect.

FIG. 6 is a claim correspondence diagram showing a third measurement when the first measurement is incorrect.

FIG. 7 is an explanatory diagram when the primary measurement is incorrect.

FIG. 8 is an explanatory diagram of a position detection device used in the position detection method of the present invention.

FIG. 9 is an explanatory diagram showing an operation shift.

FIG. 10 is an explanatory diagram showing deviation correction.

FIG. 11 is an explanatory diagram showing a final command value.

FIG. 12 is an explanatory diagram showing a conventional position detection method.

[Explanation of symbols]

1 ... Object 2 ... CCD camera 13 ... Robot a, b ... Two points c ... Line connecting two points d ... Reference line e, f ... Two points g, h ... Two points θ, θ ₂ , θ ₃ ... Angle

Claims (4)

[Claims] 1. A method for detecting the position of an object by image processing, wherein the position of two points of the object and the angle of a line connecting these two points with respect to a reference line are obtained by image recognition of visual means, and then the visual point is detected. The means is moved so that the obtained position and angle become the reference position and the reference angle, the positions and angles of the two points are detected again at this movement position, and then the distance between the visual means and the object is changed. A position detection method for correcting the position and the angle by re-detecting the position and the angle of the two points. 2. The position detecting method according to claim 1, wherein the change of the distance between the visual means and the object causes the visual means to approach the object. 3. The position detecting method according to claim 2, wherein the visual means is mounted on a robot. 4. The position detecting method according to claim 3, wherein the correction of the position and the angle corrects a control error of the robot system.