JPH0727408B2 - Robot handling device with fixed 3D vision - Google Patents

Description

Description: FIELD OF THE INVENTION The present invention calibrates the spatial offset between a visual coordinate system and a robot coordinate system when a robot is made to work using a fixed vision system. Accordingly, the present invention relates to a fixed three-dimensional visual combined use robot handling apparatus suitable for accurately handling an object having an arbitrary position / orientation using visual sense.

[Background of the Invention]

Conventionally, when the robot is used to perform work by utilizing the visual sense, a planar binary or multivalued image is recognized as the visual sense and the handling is performed based on the recognized image. Therefore, the things that can be handled were limited to those with a uniform posture. At this time, as the error between the robot coordinate system and the visual coordinate system, only the translation amount on the plane and the rotation in the plane need be considered, and the coordinate system is relatively easy to calibrate. This kind of calibration is performed in the planar vision device currently in practical use. However, as the handling work by robots increases and it becomes necessary to grasp complicated objects, it becomes necessary to recognize not only planar objects but also three-dimensional objects whose positions and orientations are not fixed. It was As a method for this purpose, a three-dimensional object recognition method using binocular vision or a range image is known. In such three-dimensional vision, spatial three-dimensional error between the robot coordinate system and the visual coordinate system is not taken into consideration, and the robot mechanism performs operations such as handling on the target object as desired. It had a problem that it could not.

[Object of the Invention]

In order to solve the above-mentioned problems of the prior art, the object of which the hand of the robot mechanism is recognized by recognizing the position and orientation of the target object on which the hand of the robot mechanism performs the handling work by the fixed three-dimensional visual means. It is an object of the present invention to provide a fixed three-dimensional vision combined robot handling device capable of performing accurate handling work on an object.

[Outline of Invention]

In order to achieve the above object, the present invention provides a robot mechanism that has a hand and drives an actuator based on a robot coordinate system to move the hand, and a robot mechanism fixed near the robot mechanism to provide a visual coordinate system. Fixed three-dimensional visual means for imaging and outputting a three-dimensional image signal in the visual coordinate system, and the position and orientation of the object in the visual coordinate system based on the three-dimensional image signal output by the fixed three-dimensional visual means An image processing means for calculating, and an actuator of the robot mechanism are driven based on information of the command position X and the posture f in the robot coordinate system in advance to operate a hand to position and control a desired position on the hand, The visual coordinate system is used by the image processing means based on a three-dimensional image signal obtained by picking up a desired portion on the hand with the fixed three-dimensional visual means. In the visual coordinate system, the position "and posture" at a predetermined position on the hand is calculated, and the information of the commanded position X and posture f in the robot coordinate system is calculated by coordinate conversion calculation into the visual coordinate system. An error calculating means in the visual coordinate system for calculating a three-dimensional error in the visual coordinate system with the information of the commanded position 'and the posture', and an image of the work object to be handled by the hand of the robot mechanism are fixed to the fixed three-dimensional. From the three-dimensional image signal obtained by imaging by the visual means, the image processing means calculates the position and orientation information of the work object calculated in the visual coordinate system by the error calculating means in the visual coordinate system. Calibrated by a three-dimensional error in the visual coordinate system, and the calibrated information on the position and orientation of the work object is used in the visual coordinate system. Robot control means for performing coordinate conversion to the robot coordinate system and drivingly controlling an actuator of the robot mechanism to control the hand to perform a handling work on the work target. It is a robot handling device with fixed 3D vision.

The supplementary description is as follows.

In order to handle an object whose position / orientation is arbitrary, it is necessary to measure the position / orientation of the object and point the robot hand in the orientation direction of the object. For that purpose, the visual coordinate system for measuring the position / orientation and the robot coordinate system in which the robot moves must be accurately calibrated. If there is a deviation between these coordinate systems, if the robot is moved based on the command value of the position / orientation and the position / orientation of the robot is visually measured, an error will occur between the two. Therefore, it is possible to use this fact in reverse, and find the mounting error when fixing the vision to the robot coordinate system from the error between the robot position / orientation command value and the visually measured position / orientation. . By using this error, the robot coordinate system and the visual coordinate system can be calibrated.

Example of Invention

 Embodiments of the present invention will be described below with reference to the drawings.

FIG. 1 is an explanatory diagram of an embodiment of a robot visual coordinate system calibration method according to the present invention, FIG. 2 is a conceptual diagram of a relative relationship between a robot coordinate system and a fixed visual coordinate system, and FIG. 3 is a fixed diagram. FIG. 4 is a functional block diagram for explaining a mounting error that occurs when the visual sense is fixed to the robot coordinate system, and FIG.

Here, 1 is a robot, 2 is a robot mount, 3 is fixed vision, 4 is an image processing device, 5 is an error calculation device, 6 is a robot control device, and 7 is a visual fixing jig.

As shown in FIG. 2, the relationship between the robot coordinate system and the fixed visual coordinate system is that the origin of the robot coordinate system is _{R 1} and each axis is the same.
_{R, R, R,} the origin of the fixed vision coordinate system _V, and the axes _{V, V,} when is _V, is expressed by the following equation. _{V = R + L x V +} L y V + L z V ... (1) Where = (L _x , L _y , L _z ) is the origin of the robot coordinate system
_A vector representing the displacement between _R and the origin _V of the fixed visual coordinate system, M _α , M _β , and M _γ are matrices representing three rotational displacements between the two coordinate systems, and as shown in FIG. 3, for example. When the rotation angles α, β, γ are taken Becomes

Now, by a command to the robot, the position in the robot coordinate system = (x, y, z) , orientation _{= (f x, f y,} f z) Moving to,
Assuming that the parameters L _x , _Ly , L _z , α, β, and γ are as designed values and do not include an error, then: = x _R + y _R + z _R + _R (4) = f _{x R} + f _{y R} It should move to + f _{z R} (5). The values calculated by the fixed visual sense 3 for this position / orientation are ′ = (x ′, y ′, z ′), ′ = (f _x ′,
f _y ′, f _z ′), ′ = x ′ _V + y ′ _V + z ′ _V + _V (6) ′ = f _x ′ _V + f _y ′ _V + f _z ′ _V (7) . x, y, z and x ', y', z 'and _{f x, f y, f z} and _{r x', f y ',} f z' relationship with the the use of the equation (1) to (3) I want it.

However, if an error occurs when the fixed vision 3 is attached, the error is the error ΔL _x , ΔL _y , ΔL such as L _x.
It is represented by _z , Δα, Δβ, and Δγ. At this time, each parameter in the equations (1) to (3) is (L _x + ΔL _x ),
Since it becomes (α + Δα) etc., the robot is calculated as = (x, y,
z), = (f _x , f _y , f _z ) is commanded to move, and it is measured by the fixed visual sense 3, the fixed visual coordinate system is' =
(X ′, y ′, z ′), ′ = (f _x ′, f _y ′, f _z ′) does not hold, and deviated values ″ = (x ″, y ″, z ″), ″ = ( f
_x ″, _fy ″, _fz ″). Therefore, x ′, y ′, z ′
And x ″, y ″, z ″, f _x ′, f _y ′, f _z ′ and f _x ″, f _y ″, f
_If each deviation of _z ″ is measured, by solving equations (1) to (3), ΔL _x , ΔL _y , ΔL _z , Δα, Δβ, Δγ
Can be asked.

The configuration and operation of this embodiment will be described with reference to FIG. 4 based on the explanatory view of FIG.

The robot 1 moves based on the position / orientation command values from the robot controller 6, (Equations (4) and (5)), and the fixed vision 3 captures the image of the robot 1 after the movement and sends it to the image processor 4. send. The image processing device 4 measures the position / orientation "," of the robot 1 viewed from the fixed vision 3. on the other hand,
The error calculation device 7 uses the formulas (1), (2), and (3) based on the command value from the robot controller 6 to calculate the position / orientation ',' (formula ( 6),
(7)) is calculated. While it is assumed that there is no parameter error when calculating ',', "," measured by the fixed vision 3 includes parameter error. Therefore, the parameter error can be obtained from each error between'and 'and'and'.

〔The invention's effect〕

As described above in detail, according to the present invention, the fixing 3
3 that occurs when the three-dimensional visual means is fixed to the robot coordinate system
Since the three-dimensional mounting error can be obtained by moving the robot, (1) the three-dimensional mounting error is obtained, and by correcting this error, not only the planar object but also the position .A three-dimensional target object having an arbitrary posture can be recognized by using fixed three-dimensional vision, and a complicated handling operation using a robot can be performed on the recognized target object. (2) Robot An object whose mounting error can be obtained by simply moving and measuring, and the calibration work is much simpler than when measuring the mounting error directly using a measuring instrument, and the position / orientation is recognized by the fixed three-dimensional visual means. It is possible to improve the accuracy and efficiency of the handling work of the robot mechanism with respect to the object.

[Brief description of drawings]

FIG. 1 is an explanatory diagram of an embodiment of a robot visual coordinate system calibration method according to the present invention, FIG. 2 is a conceptual diagram of a relative relationship between a robot coordinate system and a fixed visual coordinate system, and FIG. 3 is a fixed diagram. FIG. 4 is a conceptual diagram illustrating a mounting error that occurs when the visual sense is fixed to the robot coordinate system, and FIG. 4 is a functional block diagram of this system. 1 ... Robot, 2 ... Robot mount, 3 ... Fixed vision, 4 ... Image processing device, 5 ... Error calculation device, 6
...... Robot control device, 7 ・ ・ ・ Visual fixing jig.

 ─────────────────────────────────────────────────── ─── Continuation of the front page (72) Inventor Seiji Hata, 292 Yoshida-cho, Totsuka-ku, Yokohama-shi, Kanagawa, Ltd. Within the Institute of Industrial Science, Hitachi, Ltd. (56) Reference JP-A-57-182205 (JP, A) Kai 53-98670 (JP, A) JP 60-159905 (JP, A) Actual 57-122384 (JP, U)

Claims (1)

[Claims] 1. A robot mechanism having a hand, which drives an actuator based on the robot coordinate system to move the hand, and a robot mechanism fixed near the robot mechanism and imaged in the visual coordinate system. Fixed three-dimensional visual means for outputting a three-dimensional image signal, and image processing means for calculating the position and orientation of an object in the visual coordinate system based on the three-dimensional image signal output by the fixed three-dimensional visual means, Based on the information of the command position X and the posture f in the robot coordinate system in advance, the actuator of the robot mechanism is driven to operate the hand to position and control the desired position on the hand, and the desired position on the hand. Based on a three-dimensional image signal obtained by picking up the fixed three-dimensional visual means by the image processing means, the predetermined predetermined on the hand washing in the visual coordinate system. The position "and orientation" of the position is calculated, and the information of the command position X and orientation f in the robot coordinate system is subjected to coordinate conversion calculation to the visual coordinate system to calculate the command position'and orientation 'in the visual coordinate system. An error calculating means in the visual coordinate system for calculating a three-dimensional error in the visual coordinate system with respect to information, and an image of a work object to be handled by the hand of the robot mechanism are obtained by imaging the fixed three-dimensional visual means. To be 3
The three-dimensional error in the visual coordinate system calculated by the error calculating means in the visual coordinate system with respect to the position and orientation information of the work object calculated in the visual coordinate system by the image processing means from the three-dimensional image signal. The coordinate of the position and orientation of the calibrated work object is further converted from the visual coordinate system to the robot coordinate system to drive and control the actuator of the robot mechanism so that the hand becomes the work object. A fixed three-dimensional visual combined use robot handling device, characterized in that the robot handling means is provided with a robot control means for controlling the handling work.