JPH04100118A - Position detection for robot with visual sense - Google Patents

Abstract

PURPOSE:To control a robot to handle many kinds of work with a high precision by preliminarily indicating a feature point to be detected of a work by a reference tool and obtaining the size of the work in accordance with absolute coordinates of this feature point and detecting the feature point of the work by a camera. CONSTITUTION:The feature point to be detected of a reference work W is preliminarily indicated by a reference tool 4 attached to the front end of a robot arm 1, and absolute coordinates of this feature point are taught to a controller. The feature point of the reference work W is put in the visual field of a camera 3 to store its position in the visual field of the camera 3, and the extent of relative deviation is detected based on the position of the picture of the feature point of an actual work W' put in the visual field of the camera 3 in the actual work and the preliminarily stored position of the feature point of the reference work in the visual field of the camera. Thus, the robot is controlled to handle many kinds of work with a high precision.

Description

[Detailed Description of the Invention] [Industrial Application Field] The present invention relates to a method for detecting the position of a workpiece used in an assembly robot, etc., and particularly relates to a method for detecting the position of a workpiece used in an assembly robot, etc. The present invention relates to a position detection method for correcting the amount of positional deviation.

[Prior art] In general, image processing methods for detecting positional deviation of a workpiece using a robot equipped with a visual system such as a camera include a method called a hand-eye method, in which a camera is attached to the tip of a robot arm, and a fixed method. There is a detection method called the camera method, in which a camera is installed in a location separate from the robot.
1641).

[Problems to be Solved by the Invention] However, when it is necessary to accurately detect a measuring object with large dimensions, both methods have conflicting issues between the size of the camera field of view and the measurement resolution. Detection using a fixed camera method requires sacrificing either the size of the camera field of view or the measurement resolution.
It has the disadvantage that it is limited to applications of specific size workpieces.

In addition, in the case of the hand-eye method, by moving the camera, it is possible to accurately detect multiple feature points of the target workpiece and calculate the deviation of the jaw from the set position based on the geometric relationship. In order to do this, it is necessary to register the dimensional data of the target workpiece in advance.Also, when aligning the feature points on the camera image with the camera's reference position, it is difficult to align the position because it cannot be seen while moving the robot arm. Furthermore, in cases where the size of the workpiece is much larger than the field of view of the camera and a workpiece with a different size is accidentally supplied, it is possible to place only the vicinity of the feature point in the field of view of the camera and determine the position of the feature point. Even if detected, the position of the workpiece cannot be detected correctly unless it is equipped with a function that can check the absolute dimensions between it and other feature points.

An object of the present invention is to accurately detect the amount of deviation of a target workpiece from a set position using a hand-eye method, and also to obtain shape information indicating absolute dimensions.

[Means for Solving the Problems] The present invention provides a position detection method in which a camera is provided at the tip of a robot arm to detect the position of a workpiece. The positions of a plurality of feature points on a reference workpiece having the same shape as the workpiece are specified, the coordinate information of the specified feature points is stored, and the robot arm is moved to locate the feature points of the reference workpiece using the camera. storing a position within the camera field of view of a feature point of the actual workpiece corresponding to the feature point of the reference work; storing a position within the camera field of view of the feature point of the reference work and the actual work; Determining the amount of deviation between the feature points of the reference workpiece and the feature points of the actual workpiece from the position of the feature points of the workpiece within the camera field of view,
This is a method of determining the amount of deviation of the actual workpiece from the amount of deviation and coordinate information of the feature points of the reference work.

Further, the method is such that the coordinate information of the feature points of the actual workpiece is determined from the coordinate information of the feature points of the reference workpiece and the deviation m of the actual workpiece, and the shape information of the reference workpiece and the actual workpiece is determined. .

``Function'' Attached to the tip of the robot arm in advance, the reference tool is used to indicate the feature points to be detected on the reference workpiece, and the absolute coordinates of the feature points are taught to the control device.
The feature points of the reference workpiece are placed in the field of view of the camera and their positions within the field of view of the camera are memorized, and the positions of the images of the feature points of the actual workpiece taken into the field of view of the camera during the actual work are stored in advance. By detecting the relative deviation amount from the position of the feature point of the reference workpiece in the field of view of the camera, the reference workpiece is It is possible to obtain the absolute positional relationship between the feature points of the actual workpiece and the actual workpiece, and also to obtain shape information such as the absolute dimensions of the workpiece by calculation.

[Example] The present invention will be described with reference to embodiments shown in the drawings.

FIG. 1 is a perspective view showing an embodiment of the present invention, in which a robot arm 1 is provided with a work gripping section 2 and a camera 3 at its tip. Further, a reference tool 4 that can be attached and detached is provided so as to indicate the position of a control point that is set near the workpiece gripping part 2 and indicates the current position of the robot.

Next, we will explain the operation of detecting the position of an actual workpiece using the device configured as described above, using a reference workpiece having a shape similar to the actual workpiece to be detected in the actual work. .

First, the tip of the reference tool 4 attached to the control point position of the robot is moved to the reference position of the reference workpiece W placed at the teaching position for robot work, for example, to the position of feature points PI and P2 such as holes when tightening bolts. Move (Step 1)
.

Next, as a reference point for recognition as the absolute value of the robot coordinate system, the absolute coordinates (x, y, , x2, y2) of the feature points PI and P2 are stored as shown in FIG. 2 (step 2).

Next, the robot arm 1 is moved to place the feature point P1 in the camera field of view A of the camera 3, and the position of the feature point P1 within the field of view is memorized ((Step 3).

Next, in the actual work, the workpiece is transported in the direction of the arrow by the conveyor C shown in FIG. 1, and as shown in FIG.
In the figure, it is easy to understand that the position is greatly shifted), but the camera 3
The position of the feature point PI' corresponding to the feature point Pi is taken in as image data, and then the camera 3 is moved to the feature point P.
1 to the feature point P2, and the position of the feature point P2' corresponding to the feature point P2 is taken in as image data (step 4). Then, the feature point P from the reference position (X,, Y,, X2Y 2 ) of the feature point PL, P2 is determined by image processing from the position image data of each feature point.
1°, P2° angle 1 f11 (ΔXΔy+), (ΔX2
+Δy2) is determined (Step 5).

Next, the deviation ff1 (ΔXl + Δy+), (ΔX
2 + Δy2), the absolute coordinates (X,', YX2', Y2°) of the feature points PI' and P2° are determined by the following formula (Step 6).

x,' =x, +Δx,,Y,° =Y,+Δyx2
° =:X2-+4x2. y21 = Y2 + Δy2 Also, at the same time, the feature points Pl and P2, and the feature points P1 and P
2, the distances L1 and L2 are determined by the following calculations (step 7).

L1=q ding■ dinggo])-]q L2=-→- The deviation correction amount of the workpiece W is the deviation amount ΔX of the coordinates when the workpiece center O is shifted and moved to the position Oo.

ΔY and inclination angle θ are determined using the following equations (step 8).

ΔX=(ΔX1+ΔX2)/2 ΔY=(Δy1+Δyz)/However, L3 is the virtual feature point P when the workpiece center O is translated in parallel so that it overlaps with Oo," (X, "Yl")
and the feature point Pl゛, which is Xl”=
From X, +Δx,y,” =y, +ΔY, L 3 =
(1/2) ・X, -X, + Y, -Y.

It is required as.

FIG. 3 is a flowchart of the operation for determining the amount of shift of the workpiece.

A correction amount is obtained from the coordinate information such as the absolute coordinates of the feature points, the interval between the feature points, and the amount of shift in the coordinates of the center of the workpiece, which are obtained in the above manner, and the correction amount is given to the robot control device to correct the workpiece position.

In this way, it is possible to obtain the absolute positional relationship between the feature points of the workpiece from the absolute coordinate values of the feature points initially taught and the above-mentioned relative deviation amount, and also obtain shape information such as the absolute dimensions of the workpiece. It can be determined by calculation, and it is possible to verify differences in workpiece dimensions and even differences in models.

In addition, when capturing images of feature points on a workpiece with a camera during actual work, there is no need to pay attention to positioning the feature points to the reference points in the camera's field of view, as in the past, but simply place the feature points in the camera's field of view. All you have to do is

Additionally, when moving the camera near multiple feature points, you can easily move the camera by inputting the coordinate information already obtained from the reference work into the robot's control device.
The hassle of setting the camera position can be greatly reduced.

[Effects of the Invention] As described above, according to the present invention, the feature points to be detected on the workpiece are specified in advance by the reference tool, the size of the workpiece is determined from the absolute coordinates of the feature points, and the size of the workpiece is further detected. Since only the feature points are detected by the camera, the field of view of the camera is reduced and the resolution is increased, making it possible to detect the amount of deviation with high accuracy, and also to obtain the absolute dimensions of the workpiece even for relatively large workpieces. This has the effect of enabling robot control that handles a wide variety of workpieces with high precision.

[Brief explanation of the drawing]

FIG. 1 is a perspective view showing an embodiment of the present invention, FIG. 2 is a plane coordinate diagram illustrating the amount of deviation of a workpiece, and FIG. 3 is a flowchart. l. Robot arm, 2... Work gripping part, 3...
・Camera, 4...Reference tool, Diagram Patent applicant Yasuyo Electric Manufacturing Co., Ltd. Diagram procedural amendment (method) Indication of incident 1990 Patent application No. 172456 Name of invention Vision robot) Position detection Relationship to the case of the person amending the method Patent Applicant Address 2346 Oaza Fujita, Yahatanishi-ku, Kitakyushu City, Fukuoka Prefecture Name (662) Anyo Electric Manufacturing Co., Ltd. Figures 1 to 2 are corrected as shown in the attached sheet. Figure Figure

Claims (1)

[Claims] 1. In a position detection method in which a camera is provided at the tip of a robot arm to detect the position of a workpiece, a reference tool is attached to the tip of the robot arm, and the reference tool is set in advance to be similar to the actual workpiece. Instructing the positions of a plurality of feature points of a reference workpiece having a shape, storing coordinate information of the indicated feature points, and moving the robot arm to position the feature points of the reference workpiece within the camera field of view using the camera. , memorize the position of the feature point of the actual workpiece in the camera field of view corresponding to the feature point of the reference workpiece, and store the position of the feature point of the reference workpiece in the camera field of view and the feature point of the actual workpiece. Determining the amount of deviation between the feature points of the reference workpiece and the feature points of the actual workpiece from a position within the field of view of the camera, and determining the amount of deviation of the actual workpiece from the amount of deviation and coordinate information of the feature point of the reference workpiece. A method for detecting the position of a robot with vision, characterized in that: 2. The coordinate information of the feature points of the actual workpiece is obtained from the coordinate information of the feature points of the reference workpiece and the amount of deviation of the actual workpiece, and the shape information of the reference workpiece and the actual workpiece is obtained. The position detection method of the described robot with vision.