JPH01180401A - Correcting method for body recognition position of robot with visual sensor - Google Patents

Abstract

PURPOSE:To obtain common algorithm for correcting the position of a body recognized by an independent visual sensor and a visual sensor integrated with an arm by finding the correlation between the tip position of the arm and the body recognition position of a recognition sensor. CONSTITUTION:When the visual sensor 3 is fixed independently of the arm 10, the sensor detects the position of a mark 6 fitted atop of the arm 10 while the mark is moved together with the arm 10 and the correlation is found by the arm tip position 12 and the body recognition position of the sensor 3. When the sensor 3 is held integrally with the arm 10, the mark 6 is fixed and the sensor 3 detects the position of the mark 6 while the arm 10 is moved to find the correlation between the arm tip position 12 and the body recognition position of the sensor 3. Then the position of the body recognized by the sensor 3 is corrected into the position of the body viewed from the arm tip 12 with respective correlation functions. Consequently, when the sensor 3 is fitted to the arm, algorithm for detecting the displacement of the mark 6 by the coordinate system of a robot side is only added.

Description

[Detailed Description of the Invention] [Industrial Application Field] The present invention provides an object i\! of a robot with a visual sensor. This is a method for correcting the detected position.

[Conventional technology]

2. Description of the Related Art Conventionally, robots are known that recognize the position of an object using a visual sensor such as a television camera, and move the tip of an arm to the recognized position to perform a predetermined task such as grasping the object.

FIG. 3 is a block diagram showing the mechanical configuration of this type of robot. The b-bot 1 is equipped with two arms 10 and 11, and the three-dimensional position of the tip of each arm is controlled by a controller 2. A gripping member 12 for gripping an object is attached to the tip of the arm 10 at the working end. On the other hand, a visual sensor 3 consisting of a television camera is connected to the controller 2, which recognizes the position of an object 5 placed on the workbench 4 based on the output image 1 of the visual sensor 3. It is configured to move the gripping member 12 at the tip of the arm 10 to a position to grip the object 5 and transfer it to, for example, another workbench.

In such a case where the position of the robot 1 and the position of the visual sensor 3 are different, it is necessary to correct the position of the object 5 recognized by the visual sensor 3 to the position of the object 5 from the grasping member 12 of the robot 1. That is, since the positions of the viewpoint and the work point are different, in order to grasp the object 5, it is necessary to convert the position of the object seen from the viewpoint to the position seen from the work point.

Therefore, in such robots, the following correction method is implemented in which the position of the object as seen from the viewpoint is corrected to the position of the object as seen from the work point. That is, as shown in FIG. 4, a mark 6 for position correction is attached close to the gripping member 12, and in this state, the arm 10 is moved two-dimensionally or three-dimensionally.
The visual sensor 3 is made to recognize the position of a plurality of sample points while displacing them in the dimensional direction. That is, in the 70-chart of FIG. 5, (XO, Yo, 20), (Xl,
Yl, Zl), (X2, Y2, Z2
) - ... (Xn, , Yn, Zn) A command to displace the sample point to a plurality of sample point positions is output from the controller 2, and the arm 10 is displaced to these sample point positions.

During this displacement process, the position of the mark 6 is recognized using the coordinate system U and ■ of the visual sensor 3 as shown in FIG. After this, each displacement position of the arm (Xi).

”y'i, Zi) and the recognition position of the visual sensor 3 (u
i.

Vi) to find the correlation function (X, Y, Z) = f (U, V) between both positions, and when actually grasping the object 5, use this correlation function to determine the position of the object 5 recognized by the visual sensor 3. After correction, a displacement command for the arm 10 is output.

Thereby, the tip of the arm 10 can be accurately displaced while the position of the object 5 is recognized by the visual sensor 3.

[Problem to be solved by the invention]

By the way, in the above-described robot, since the visual sensor 3 is fixed, the tip of the arm cannot be changed to a range that would deviate from the visual field. Therefore, it is desirable that the visual sensor 3 be configured to change integrally with the tip of the arm.

However, when the visual sensor 3 is attached to the tip of the arm,
It becomes necessary to design an algorithm for correcting the position of the object recognized by the visual sensor 3 separately from the algorithm used when the visual sensor 3 is made independent. Therefore, the structure of the algorithm becomes fundamentally different depending on whether the visual sensor 3 is independent or integrated, leading to a problem of increased cost.

The purpose of the present invention is to have an independent visual sensor and an integrated structure in the arm! i1! An object of the present invention is to provide a method for correcting the object recognition position of a robot with a visual sensor, which allows a common algorithm for correcting the position of an object that has been moved.

(Means for Solving the Problems) The present invention provides that, when the visual sensor is fixed independently from the arm, the visual sensor detects the position of the mark attached to the tip of the arm while moving it together with the arm. The correlation function between the position and the object recognition position of the visual sensor is calculated, and when the visual sensor is held integrally with the arm, the mark is fixed, and while the arm is moved, the mark is detected by the visual sensor held on the arm. is configured to detect the position of the arm, find a correlation function between the arm tip position and the object recognition position of the visual sensor, and use this correlation function to correct the position of the object recognized by the visual sensor to the position of the object as seen from the arm tip. This is what I did.

[Effect]

When the visual sensor is made independent from the arm, the position of the object recognized by the visual sensor is corrected in the same way as in the past. On the other hand, when the visual sensor is integrated with the arm, it is assumed that the mark is in the first position when the arm is displaced, and the correlation function between the arm tip position and the object recognition position of the visual sensor is calculated. The position of the object recognized by the visual sensor is corrected by

Therefore, when the visual sensor is attached to the arm, it is only necessary to add an algorithm to detect the displacement of the mark using the robot's coordinate system.

〔Example〕

FIG. 1 is a flowchart showing an embodiment of the present invention, in which steps 85 to S9 are added to the conventional procedure shown in FIG. Therefore, only steps 85 to S9 will be explained. First, in step S5, it is checked whether the visual sensor is integrally attached to the arm of the robot. That is, as shown in FIG. 2 (a), it is checked whether the visual sensor 3 is attached to the tip of the arm 10 of the robot 1.Next, if the visual sensor 3 is attached to the tip of the arm 1o, For example, the tip of the arm is moved to the position where the target object is actually recognized.
Obtain (Xo, Yo, Zo). Manually displace the gripping member 12 at the tip of the arm, align it with the center of the mark 6 as shown in FIG. 2(b), and turn it on.

Yi, Z+u, then the arm tip is returned to the reference position (a), and the visual sensor 3 is made to recognize the position of the mark 6 in this state. Next, in steps 87 to S9, arm 1
0 tips (Xl, Yl.

Zl), · (Xn, Yn, Zn) commands to displace the sample points to a plurality of positions are output, and the arm 10 is displaced to these sample point positions. Then, as shown in FIG. 2(C), even though the visual sensor 3 has been displaced with respect to the mark 6, the visual sensor 3 sees that the mark 6 has been displaced. The coordinate position is (Xo-Xi) +Xn, (Yo-Yi
) +Yn+.

(Zo-Zi) +Z+a (However, i=1.2・
). Therefore, the mark 6 at this coordinate position is recognized in the U, ■ coordinate system of the visual sensor 3, and the opening is determined by the recognition position (Ui, Vi) at each sample point position and the displacement position (Xi, Yi, Zi) of the mark 6. Find the position correlation function (X, Y, Z) = f (U, V).

In this way, even when the visual sensor 3 is integrally attached to the arm 10 of the robot, the position of the object seen from the viewpoint can be corrected to the position seen from the work point in the same way as when it is not attached.

Therefore, by simply adding the step S8 of detecting the position of the mark 6 as if the mark itself has been displaced to the conventional procedure, it is possible to configure the same algorithm as the conventional one.

[Effects of the Invention] As explained above, according to the present invention, when the visual sensor is integrated into the arm of the robot, it is assumed that the mark is displaced in response to the displacement of the arm, and the object recognized by the visual sensor is Since the position is corrected, a common algorithm can be used for those in which the visual sensor is integrated into the arm and those in which the visual sensor is independent from the arm.

[Brief explanation of the drawing]

FIG. 1 is a flowchart showing one embodiment of the present invention, and FIG.
The figure is an explanatory diagram of the process of correcting the object position as seen from the viewpoint, FIG. 3 is a schematic configuration diagram showing an example of a conventional robot with a visual sensor, and FIG. 4 is an explanatory diagram of the conventional method of correcting the object recognition position. Figure 5 is a flowchart showing the procedure, and Figure 6 is a flowchart showing the procedure.
The figure is an explanatory diagram showing the recognition position of an object as seen from a visual sensor. 1... Robot, 2... Controller, 3... Visual sensor, 5... Object, 6... Mark, 10.11
···arm. (Q) (b) Xo-Xi Yo -'V+ Zo -Zi (C) Figure 3 Figure 4 Figure 5 Figure 6

Claims (1)

[Claims] In a robot with a visual sensor that recognizes the position of an object using a visual sensor and moves the tip of the arm to the recognized position to perform a predetermined task, if the visual sensor is fixed independently from the arm, the tip of the arm The position of the attached mark is detected by the vision sensor while moving with the arm, and the correlation function between the arm tip position and the object recognition position of the vision sensor is determined, and when the vision sensor is held integrally with the arm, the mark is fixed. Then, while moving the arm, the position of the mark is detected by the visual sensor held on the arm, and a correlation function between the arm tip position and the object recognition position of the visual sensor is determined, and the visual sensor recognizes the object based on this correlation function. A method for correcting an object recognition position of a robot with a visual sensor, the method comprising: correcting the position of the object seen from the tip of the arm.