JPS61270090A - Method of controlling robot with camera for visual recognition - Google Patents

Abstract

(57) [Summary] This bulletin contains application data before electronic filing, so abstract data is not recorded.

Description

DETAILED DESCRIPTION OF THE INVENTION Field of Industrial Application The present invention relates to a method of controlling a robot (hereinafter referred to as a robot) having a visual recognition camera (hereinafter referred to as a camera) attached to a part of its arm.

BACKGROUND ART Generally, when a camera is attached to a robot, the camera 3 and visual recognition device 4 are installed in a fixed position, and the robot arm 1 is physically separated from the robot control device 2, as shown in FIG. was.

Problems to be Solved by the Invention In the case of the above conventional example, the visual recognition range is fixed, and if the field of view is widened, the resolution decreases, and if multiple visual recognition points are required, multiple cameras are required. In addition, the camera support stand and the like are complicated, which limits the robot's movable range.

Means for Solving the Problems In order to solve the above problems, the present invention, as shown in FIG. While changing the recognition range 9 magnification and direction, the presence or absence and position of the recognition mark is searched in two reference areas within the movement range of the robot arm, and the position between the two recognized recognition marks is searched. Compare the distance with the theoretical value, which is the physical dimension of the recognition target, and if the difference between the two values is within a predetermined range,
Furthermore, a correction value is calculated from the difference between the reference position of the recognition mark registered in advance and the position of the recognized recognition mark, and the working position of the robot arm is corrected using the correction value. The present invention provides a control method.

Operation The present invention uses the above means to attach a camera to a part of the robot's arm, and has a recognition range of 9 magnification in conjunction with the movement and stop of the robot's arm, a function to perform recognition search while changing the direction, and a function to perform a recognition search while changing the direction. It has a function to evaluate the validity of the recognition result by comparing the calculated value with the theoretical value, which is the physical dimensions of the recognition target, and a function to calculate the correction value based on the recognition result and correct the robot's working position. This makes it possible to widen the recognition range and make it more versatile.

EXAMPLE Hereinafter, an example of the present invention will be described with reference to the drawings. FIG. 4 is a block diagram of the embodiment,
8 is a gripping mechanism attached to the robot body, 9 is a carrier, 1o is a recognition mark on the carrier, and 11 is a workpiece (hereinafter referred to as a work) on the carrier. In the example, the workpiece is 1
Several conveyors move close to the robot and stop, but the stopping accuracy is not good, so two recognition marks placed on the conveyor are imaged with the camera 3 attached to the robot's arm. The information is input and recognized by the visual recognition device 4, and positional deviation and rotation of the transport body are detected. One to several workpieces are placed on the conveyor while maintaining a constant distance from the recognition mark, and the robot control device 2 grasps the recognized deviation amount of 2 rotations and determines the gripping position of the workpiece. After the correction calculation, it is gripped by the gripping mechanism 8.

In addition, in the embodiment, the positioning accuracy of the carrier is relatively good in the direction perpendicular to the movement direction and the plane (hereinafter referred to as the y direction), but not so good in the movement direction (hereinafter referred to as the X direction). It may not go inside. Therefore, in the embodiment, a plurality of recognition areas 12 (
The image input range (range in which images can be input) can be set with one camera operation, and the presence/absence and position of a recognition mark is recognized by performing a recognition search and its evaluation in accordance with the procedure shown in FIG. Although the recognition areas can be set arbitrarily, in the embodiment, each recognition area is provided with an overlap portion 13 to prevent recognition failure on the boundaries of the recognition areas.

Furthermore, in the embodiment, the physical distance (L) between the recognition marks
Since L' is constant, the distance (L') obtained from the recognition result is evaluated using the following equation to prevent misrecognition.

(1-α)L≦L'≦(1+α)L Here α is a coefficient.

Also, from Fig. 7, L' is calculated by the following formula: L' = [(
x2-xl +α3"2-xm)-α1("1-"m)
)20(Y2-Y1+α4(y2-ym)-α2(7
1-7m month"2]%Here P(Xl, Yl)...Position of the robot in recognition area 1 P(X2.Y2)...Position of the robot in recognition area 2 P( ``m'' engineering)...... Reference position (pixel value) when inputting an image P (xl, yl)...... Position visually recognized in recognition area 1 (pixel value) α1.α2 ......One direction in recognition area 1.

Magnification P in the y direction (x2.y2)...Position visually recognized in recognition area 2 (pixel value) α3. α4......X direction in recognition area 2.

Magnification in y direction It is.

Next, regarding the correction of the workpiece gripping position in the example,
This will be explained using FIG. PMl the reference position of the recognition mark
, 2M2, and the recognized mark positions are PRl and PH1.
Then, the workpiece gripping position is corrected as follows.

■ Amount of parallel movement of the conveyor (ΔX) from PMl and PRl.

Find Δy).

■ "Ml" Find the rotation amount θ from the slope of the line segment of M2 and the line segments of PRl and PH1.

■ ■、■Use the value obtained by Eri and set the reference position to Δ”+'
Translate by V and rotate by θ around R1.

In this way, by attaching a camera to a part of the robot's arm and adding recognition search, recognition verification, and recognition correction functions to the robot control device, we have created a system that can perform recognition and correction work throughout the robot's range of motion. can be easily constructed.

Effects of the Invention As described above, in the present invention, a positioning regulating device for the entire workpiece within the movable range of the robot can be omitted or simplified, so that a highly flexible system can be realized at low cost.

[Brief explanation of the drawing]

FIG. 1 is a perspective view showing a conventional example, FIGS. 2 and 4 are perspective views and block diagrams of an embodiment of the present invention, FIG. 3 is a flowchart showing the basic procedure of the present invention, and FIG. Fig. 5 is an explanatory diagram of the recognition search method, Fig. 6 is a flowchart of the recognition search, Fig. 7 is a related diagram of the position where the distance is determined based on the recognition result, and Fig. 8 is the correction of the work gripping position in the same embodiment. It is a related diagram of a procedure. 6...Cable between robot control device and visual recognition device, θ...Cable between camera and visual recognition device, 7...Camera support column, 12...・・・
- Recognition area, 13...Overlapping part of recognition areas. Name of agent: Patent attorney Toshio Nakao and 1 other person] Figure 1 Figure 2 Figure 3 Fo”°-1 M Jh Mark 13-1-SR-IC-R,・115 Figure 6

Claims (1)

[Claims] A visual recognition camera is attached to a part of the robot's arm,
The range of recognition changes in conjunction with the movement and stopping of the robot arm.
While changing the magnification and direction, the existence and position of recognition marks are recognized and searched in two reference areas within the movement range of the robot arm, and the distance between the two recognized recognition marks and the recognition target are searched. is compared with the theoretical value, which is the physical dimension of A control method for a robot equipped with a camera for visual recognition, which calculates a correction value and corrects the working position of a robot arm using the correction value.