JP2000199704A - Calibration method for image processor - Google Patents

Abstract

PROBLEM TO BE SOLVED: To attain with high accuracy the calibration of an image processor and a moving table, when an error at the time of calculating the position of the center of rotation is made large due to the impossibility of the increase of the angle of rotation of a θ axis of the moving table. SOLUTION: The specific pattern of a calibration mark or work 10 at a first point is recognized from the photographed image of the calibration mark or work 10 placed on an XYθ table 14, and the position (x1, y1) is measured, and the θ table is rotated so that the calibration mark or work 10 can be deviated from the range of the field of sight of a camera 11. Then, the X or Y table is moved within the field of view of the camera, and the specific pattern of the calibration mark or work 10 at the second point is recognized, and the position (x2, y2) is measured, and the position of the center of rotation of the θtable is calculated, based on position information (tx1, ty1, tθ1) of the XYθ table at the first point and position information (tx2, ty2, tθ2) of the XYθtable at the second point.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

[0001] 1. Field of the Invention [0002] The present invention relates to an image processing apparatus for recognizing the position of a work, and a method for calibrating the positional relationship between a moving table (XYθ table) for moving the position of the work.

[0002]

2. Description of the Related Art In an automation device using an image processing apparatus, it is necessary to calibrate a positional relationship between a moving table for moving a work to be worked by the automation device and a camera for photographing the work. The moving table generally includes an X axis, a Y axis, and a θ axis on the X axis and the Y axis for rotating the work. For the calibration of the pixel size, which is the basic resolution of the image processing apparatus, only one of the X-axis and the Y-axis of the moving table is used, and a calibration mark or a work recognizable by the image processing apparatus is moved to the camera. Any two points in the field of view are moved, the calibration mark or the work is recognized by the image processing apparatus, and the position is calculated from the position data and the moving distance of the axis (X axis or Y axis) of the actual stage. Further, in order to perform the rotation correction based on the position data of the work measured by the image processing apparatus using the moving table having the θ axis, it is necessary to find the rotation center position of the θ axis. Conventionally, calibration marks or workpieces recognizable by the image processing device are rotated by rotating the θ axis at any two points in the camera's field of view, and the positions are measured by the image processing device. A method of obtaining the rotation center position from the movement amount of the θ-axis based on the later recognized position data has been adopted. FIG. 8 is a view of the XYθ table 14 as viewed from above, and shows a conventional calibration method in which the θ table is rotated with the calibration mark or the actual work 71 in the field of view of the camera.

[0003]

As the number of automated devices requiring fine positioning accuracy is increasing, the magnification of a camera is increasing and the field of view is becoming narrower. In order to move (rotate) two points in the narrow field of view and obtain the position by the image processing apparatus, the rotation angle of the θ axis of the moving table cannot be increased, so that an error in calculating the rotation center position by calculation becomes large. There is. It is an object of the present invention to improve such a problem and to calibrate an image processing apparatus and a moving table with high accuracy.

[0004]

In order to achieve this object, the present invention relates to an image processing apparatus for recognizing a position of a work, wherein a calibration mark or a work placed on an XYθ table is photographed and photographed. The first point of the calibration mark or the specific pattern of the work is recognized from the image, the position (x1, y1) is measured, and the θ table is set so that the calibration mark or the work protrudes from the range of the field of view of the camera. Then, the X table or the Y table is moved into the field of view of the camera, the second calibration mark or the specific pattern of the work is recognized, and the position (x2, y2) is measured. The position information (tx1, ty1, tθ1) of the XYθ table of the first point and the second
The position information (tx2, ty2, t
The rotation center position of the θ table is obtained based on θ2).

[0005]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS An embodiment of the present invention will be described below in detail with reference to the drawings. FIG. 1 shows an example of a configuration for implementing the present invention, in which 11 is a television camera, 12 is illumination, 13 is an image processing device, 14 is an XYθ table,
5 is a machine controller. Table 14 has a calibration mark or an actual work on the XYθ table 14, and includes an X table, a Y table, and a θ table. The image processing device 13 processes a signal from the television camera 11 and outputs the signal to the machine controller 15. The machine controller 15 controls the driving of the XYθ table 14. The illumination 12 is provided so that the image processing apparatus 13 can easily recognize the calibration mark or the actual work on the XYθ table 14. Figure 2 shows the XYθ table 1
4 is a top view of the first calibration mark or the actual work 21 on the table.
Is arranged. FIG. 3 is a view of the XYθ table 14 as viewed from above, and shows the calibration mark or the actual work 31 after the θ table of the XYθ table has been rotated so large that it protrudes from the camera field of view. FIG. 4 is a view of the XYθ table 14 as viewed from above, and the XYθ table 14 is moved out of the camera field of view due to the rotation of the θ table. The second calibration mark or the actual work 41 is shown. FIG. 5 is a diagram comparing the camera field of view according to the present invention with the prior art.
The rotational position can be obtained by calculating the isosceles triangle based on the length (W) of the base and the angle (θ) at which the isosceles faces.
The difference between the prior art and the present invention is clear when the height error (εw) and the angle error (εθ) are the same as when the height (H) of the equilateral triangle is determined. As is clear from the figure, H = (W1 / 2 ± εw) / tan (θ1 / 2 ± εθ) in the prior art H = (W2 ± εw) / tan (θ2 / 2 ± εθ) , W1 <W2, θ1 <θ2 Therefore, the larger W and θ, the smaller the variation of H, and the rotation center position can be obtained with high accuracy. FIG. 6 shows FIG.
5 shows the coordinate system of the image processing apparatus. FIG. 7 is a diagram in which the coordinate system of the image processing apparatus is converted into an XYθ table coordinate system. In FIG. 6, dx = tx2−tx1, dy =
ty2-ty1, which indicates the difference between the position of the XYθ table when T1 is recognized and the position of the XYθ table when T2 is recognized, that is, the amount of movement of the table. A method for obtaining the θ-axis rotation center will be described below. 1) Two points T1 (tx1, ty) recognized by the image processing device
1), T2 (tx2, ty2) is converted from the image processing coordinate system into XY
θ table coordinate system T1 ′ (tx1 ′, ty1 ′), T2 ′
(tx2 ', ty2'). T1 'is tx1' = (tx1-Tox) * cosθ + (ty1-Toy) * sinθ ty1 '= (ty1-Toy) * cosθ-(tx1-Tox) * sinθ T2' is tx2 '= (tx2-Tox) * cosθ + (ty2-Toy) * sinθ ty2 '= (ty2-Toy) * cosθ-(tx2-Tox) * sinθ 2) From the XYθ table movement amount, the actual coordinate T2''after the movement of T2' is T2 ' Is determined based on Tx2 ″ = tx2 ′ + dx Ty2 ″ = ty2 ′ + dy 3) Obtain the rotation center of two points (T1 ′, T2 ″) in the XYθ table coordinate system. The rotation angle β of the θ table from the first point to the second point is β = Tθ2-Tθ1 (tx2 ''-xθ) = (tx1 '-xθ) * cosβ-(ty1'-y
θ) * sinβ (ty2 ''-yθ) = (ty1 '-yθ) * cosβ + (tx1'-x
θ) * sinβ Rotation center position xθ = (tx2 '' + tx1 ') / 2-(ty2''-ty1') / (2 * t
an (β / 2)) yθ = (ty2 '' + ty1 ') / 2 + (tx2''-tx1') / (2 * t
an (β / 2)).

[0006]

As described above, the calibration method of the present invention is not limited to the camera field of view,
There is an effect that the θ table can be rotated over a wide range and the rotation center position of the θ table for performing position correction with high accuracy can be obtained.

[Brief description of the drawings]

FIG. 1 is a configuration diagram of an embodiment of the present invention.

FIG. 2 is a diagram showing a first calibration mark when the XYθ table is viewed from above.

FIG. 3 is a view showing a calibration mark obtained by rotating the θ table when the XYθ table is viewed from above.

FIG. 4 shows the calibration mark shown in FIG.
The figure which shows the calibration mark of the 2nd point which moved the table or the Y table into the camera visual field.

FIG. 5 is a diagram comparing the camera field of view according to the present invention with the prior art.

FIG. 6 is a view for explaining an image processing apparatus coordinate system of the calibration mark shown in FIG. 4;

FIG. 7 is a diagram converted into an XYθ table coordinate system.

FIG. 8 is a diagram illustrating a conventional technique, and is a diagram illustrating a calibration mark rotated from a first point to a second point in a camera field of view.

[Explanation of symbols]

 11 TV Camera 12 Lighting 13 Image Processing Device 14 XYθ Table 15 Machine Controller

Claims (1)

[Claims] In an image processing apparatus for recognizing a position of a work, a calibration mark or a work placed on an XYθ table is photographed, and a first calibration mark or a specific work of the work is specified from the photographed image. The pattern is recognized, its position (x1, y1) is measured, the θ table is rotated to the extent that the calibration mark or the work protrudes from the range of the field of view of the camera, and then the X table or the Y table is placed in the field of view of the camera. Move to recognize the second calibration mark or the specific pattern of the work, and recognize the position (x
2, y2), and the position information (tx1, ty1, tθ1) of the XYθ table at the first point and the position information (tx2, ty2, tθ) of the XYθ table at the second point.
A calibration method for an image processing apparatus, wherein the θ table rotation center position is obtained based on 2).