JPH0622720B2 - Automatic coating device using teaching robot - Google Patents

Description

Description: FIELD OF THE INVENTION The present invention relates to an automatic applicator of the kind that can be used to apply adhesives and the like to automotive window glass, and in particular to TVs.
The present invention relates to a visual device using a camera and an automatic coating device using a teaching robot.

(Prior Art) Japanese Patent Application Laid-Open No. 59-498 discloses such an automatic coating device.
The one shown in Japanese Patent Publication No. 69 is known. This product deals with the positional deviation of the work by mechanical copying, but it is not suitable for high speed work because the shape and configuration of the work and jig are restricted due to the mechanical system that constitutes the copying means. There was a problem. In general, a window glass for an automobile has two sides which are substantially straight lines, two other sides which are curved lines, and a corner portion which is also a curved line, as shown in FIG. As a method of detecting the position shift of the work, a method of obtaining the center of gravity and the inclination of and can be considered, but it is difficult to detect because the work itself is large and almost transparent, and even if it can be detected, the accuracy is low. Therefore, it is considered effective to measure the amount of positional deviation at each corner (fourth corner). However, since the shape of the corner portion is also composed of curves, there is a problem that it is difficult to define the corner point like a quadrangle.

(Problems to be Solved by the Invention) An object of the present invention is to provide a visual device using a TV camera and an automatic coating device using a teaching robot, which can detect a positional deviation of a work at high speed and therefore can perform high speed work. Especially.

(Means for Solving the Problems) Therefore, the present invention solves the above-mentioned problems of the prior art by providing an automatic coating device using the teaching robot described in the claims.

(Example) Next, an example of the present invention will be described with reference to the drawings.
An automatic coating device using the teaching robot is shown in FIG.
As shown in FIG. 1B, a visual device (10) and a teaching robot (20)
In the embodiment, the visual device (10) is composed of four TV cameras (1) showing four corners of the work (11) and a TV.
A / D converter (2) that outputs the data input from the camera (1) with a value according to the brightness of each digital pixel, and A
An image memory (3) capable of storing a digital value corresponding to the brightness of each pixel output from the D / D converter (2) and storing all pixels for one screen, and the image memory (3) are arbitrarily Includes read / write hardware and software.
Further, the visual device (10) is a first straight line on the output image of the TV camera (1) of FIG. 3 showing an image of a corner portion of the work (11) shown by one of the TV cameras shown in FIG. 1B. A plurality of first window devices arranged to detect (12) (14)
And detecting the first straight line (12) as the first virtual line (17) by scanning the first window device (14) in the first direction.
Scanning device (4), a plurality of second window devices (15) arranged to detect the second straight line (13) on the output image of the TV camera (1) of FIG. 3, and Window device of 2 (15)
Scan in the second direction and make the second straight line (13) the second virtual line.
Second scanning device (5) detected as (18) and each scanning device
A virtual corner point detection device (6) for detecting by calculation the intersection (16) of the first and second virtual lines (17) and (18) detected by (4) and (5) as a virtual corner point (16) And the image memory (3) in advance
The first straight line and the second straight line of the masterwork input to
The position deviation amount detection unit (7) that detects the position deviation amount between the master work virtual corner point and the virtual corner point position, which is the intersection position of the straight line on the TV camera image, and converts this to the position deviation amount of the robot. And. Further robot
(20) has a coordinate conversion unit (9) for correcting the position deviation amount of the work (11) by converting the teaching position into coordinates based on the position deviation amount. Also, the visual device (10) is the work
The outer dimensions of the work are found by finding the four virtual corner points (16) of (11), and the robot has a work part number detection unit (9 ') that detects the work part number from the outer dimensions of the work, and the robot ( 20) has a work program selection unit (23) for inputting a work model number signal from the work model number detection unit (9 ') and selecting a plurality of types of coating work programs stored in advance by the work model number signal. Robots (20)
Has a control unit (21) and a manipulator (22).

As shown by the work (11) in FIG. 2, in general, an automobile window glass has two sides which are straight lines and the other two sides which are curved lines, and the corner portions are also formed by curved lines. These features make it difficult to define corner points. The present invention detects two different points on each side based on the above characteristics, considers a straight line formed by the two points as a virtual line, and considers a position where the virtual line intersects as a virtual corner point. By calculating the deviation amount of this virtual corner point, the positional deviation of the work (11) having a shape like a window glass for an automobile can be detected, and at the same time, the work external dimension can be obtained from the virtual corner point, and the work model number can be calculated from this external dimension. Detect and input the work model number and work position deviation amount to an automatic coating device such as a robot through a communication line (RB232C), etc., and the automatic coating device selects a work program based on the work model number signal and automatically detects the work position deviation amount. The automatic coating device is characterized in that the work is performed by correcting the locus. That is, the visual device (10) of the present invention is a TV camera based on the image of the corner portion of FIG. 3 showing the image of the corner portion of the work (11) shown by one of the TV cameras shown in FIG. 1B. As shown in FIG. 4, the two first window devices (14) and the first window device (14) are scanned in a substantially horizontal direction, and the first straight line (12) is converted into a first virtual line. The first scanning device (4) which is first detected as (17), and the two second window devices (15) for the TV camera as shown in FIG.
And scan the second window device (15) in a direction parallel to the first imaginary line (17) to move the second straight line (13) to the second imaginary line.
A second scanning device (5) for detecting as (18) is provided. Further, the visual device (10) has first and second virtual lines (17) detected by the scanning devices (4) and (5), as shown in FIG.
The virtual corner point detection device (6) that detects the intersection point (16) of (18) as a virtual corner point (16) by calculation, and the work (11) that is the teaching intersection point position previously input to the image memory (3). Positional shift amount for detecting the positional shift amount between the intersection point position of the first straight line (12) and the second straight line (13) on the TV camera image and the virtual corner point position and converting this to the positional shift amount of the robot And a detector (7).

In operation, assuming that the corners of the window glass are imaged as shown in FIG. 3, first, the first tracking is performed to find the straight lines (12) and (13) that are the sides formed by the straight lines of the window glass. . That is, as shown in FIG. 4 (a), two preset first windows (small measurement windows) (14) are moved along the preset tracking direction. While moving, it always calculates the area and center of gravity in the window, and when it exceeds a certain value, it is judged that the side of the window glass has been detected.
The position of its center of gravity is detected. (Fig. 4 (b)) The two points thus obtained give the position and inclination of the straight line portion of the window glass on the screen, and this is the first virtual line (17) (Fig. 4).
(c)). Then, a second trace is performed to detect the side composed of the curve. The second trace is a straight line with the same slope as the first virtual line (17) obtained by the first trace, which is moved by the preset offset amount 1 and offset amount 2 from the first virtual line (17). Follow (Fig. 5 (a)). This is to prevent the virtual line obtained from being different for each detection because the position to be detected is different when the position shift occurs when the side to be detected is a curve. This allows you to get two points on the side (13) made up of the curved window glass,
The straight line composed of these two points is considered as the second virtual line (18) (Fig. 5 (c)). Next, the first and second virtual lines (17), (18)
Then, the intersection is obtained and it is set as a virtual corner point (16) (6th
Figure). If virtual corner points are measured and stored for the master work, the deviation amount can be calculated from the difference from the master position even if the deviation occurs during the actual work. The virtual corner points are obtained by the above method, and the deviation amounts are output to an automatic coating device such as a robot via a communication line (RS232C or the like). The automatic coating device performs the coating work by converting the coordinates of each teaching point based on each shift amount. As a result, even if the work is misaligned, the work can be corrected.

Further, the outer dimensions of the work can be obtained by finding the four virtual corner points of the work by the above method. A visual device discriminates the work model number from the outer dimensions and informs the automatic coating device of the automatic coating device. The automatic coating device can select a plurality of types of coating work programs stored in advance by the workpiece model number signal.

(Effects of the Invention) As described above, in the present invention, the virtual corner point at the corner portion of the work is detected, and the positional deviation of the work is detected by the virtual corner point. Since the deviation amount of the work is measured and the teaching robot corrects the teaching position based on the deviation amount, it is suitable for high-speed work, and the structure of the injection nozzle system for coating can be simplified.

That is, according to the present invention, a jig having a rough accuracy can be dealt with even though it takes a lot of design time and cost to improve the accuracy of a conventional jig. Furthermore, since there is no mechanical copying device, the injection nozzle system for coating can be simplified and lightened. This simplifies the teaching of the robot and realizes a good operation. Further, since the positional deviation of the work is detected in advance before the work and the coordinate conversion of the teaching point is performed, the work can be performed at a higher speed than the mechanical copying. Selectively,
In this automatic coating device, the TV camera is installed only in the upper direction of the work and only the two-dimensional misalignment amount is measured. However, when the position deviation occurs in the vertical direction, the camera can be installed in the horizontal direction. It is also possible to detect positional deviation in three dimensions. Further, since the present apparatus of the embodiment uses the visual device, it is possible to add not only the position deviation detection but also the function of model number discrimination by measuring the dimension thereof.

[Brief description of drawings]

FIG. 1A is a schematic block diagram showing the configuration of an automatic coating apparatus using a teaching robot which is an embodiment of the present invention, and FIG. 1B is a perspective view showing the schematic hardware of the apparatus shown in FIG. 1A.
The figure shows a plan view of the work shown in FIG. 1B, and FIG. 3 shows an image of a corner portion of the work on an output image projected by one of the TV cameras shown in FIG. 1B, and FIG. First window device 2 of the visual device of FIG. 1A based on the image of FIG.
Sequentially perform the work of detecting the first virtual line by (a) (b)
(c) An explanatory view shown in each drawing, and FIG. 5 sequentially shows the work of detecting the second virtual line by two second window devices similar to FIG. 4 (a) (b) (c). Explanatory drawing shown in each figure, FIG.
It is explanatory drawing which shows the operation | movement of the virtual corner point detection apparatus which detects the virtual corner point which the intersection of the detected 1st and 2nd virtual line shown in FIG.6 (c) forms. 1 ... TV camera 4, 5 ... Scanning device 12, 13 ... Side straight line 14, 15 ... Window device 16 ... Virtual corner point 17, 18 ... Virtual line

Claims (2)

[Claims] 1. A TV camera for outputting an image of a corner portion including a first straight line of a work having a corner portion, a second straight line and a curve connecting the straight lines in an automatic coating apparatus using a teaching robot. A plurality of first lines arranged to detect the first straight line on the output image of the TV camera
Window device and the first window device
And a plurality of first scanning devices arranged to detect the second straight line on the output image of the TV camera by scanning the first straight line as a first virtual line. Second window device, a second scanning device which scans the second window device in a second direction and detects the second straight line as a second virtual line, and the second scanning device which is detected by each scanning device. A virtual corner point detection device for detecting the intersection of the first and second virtual lines as a virtual corner point by calculation, and the first part of the master work previously input to the image memory.
Of the master work virtual corner point, which is the position of the intersection of the straight line and the second straight line on the TV camera image, and the virtual corner point position, and converts this to the position shift amount of the robot. An automatic coating apparatus using a teaching robot, comprising: a detection unit; and a coordinate conversion unit that corrects the position displacement amount of the work by performing coordinate conversion of a taught position based on the position displacement amount. 2. A work model number detecting unit for determining a work model number from the work dimension by determining virtual corner points of each corner of the work and determining a distance between the virtual corner points. The automatic coating device using a teaching robot according to claim 1, further comprising: a work program selection unit that selects a coating work program stored in advance according to a work model number signal from the work model number detection unit.