JPH04134202A - Visual coordinate, object coordinate system calibration method - Google Patents

Abstract

PURPOSE:To convert positional information between a visual sensor coordinate system and an object coordinate system by taking the two dimensional position information of a fixture placed inside TV camera vision into a memory of a picture image processing device to obtain the centers of gravity of the four holes in the fixture and using the distances among the centers. CONSTITUTION:Fixture picture image is memorized in the memory of a picture image processing device 6 through vision 2. Center of gravity of two fixture holes which face x-axis and y-axis of the vision coordinate are placed so as to be in parallel to each other. With this constitution, the center of gravity is calculated by the picture image processing device by using a known algorithm, indicated in the respective holes on a TV monitor, and the fixture is moved so that the x-axis and the y-axis of the facing holes may be equal. The distance between the holes which are parallel to the x-axis and y-axis are obtained respectively. Besides, the value is substituted in the following formula. X = xcosbeta-zsinalpha Y = xsinalphasinbeta+ycosalpha-zsinalphasinbeta Z = -xsinbeta-cosalpha+ysinalpha+zcosalphacosbetaz = 0DELTAX(known distance between holes of fixture) = (x center of gravity of hole 4 - x center of gravity of hole 3). beta is determined from cosbeta. 'alpha' is determined by using the formula in the same manner. Accordingly '' for rotational angle of the coordinate axis can be determined. Besides, it is thus possible to calibrate the arbitrary coordinate value of the visual coordinate system into an objective coordinate value.

Description

[Detailed Description of the Invention] [Objective of the Invention] (Industrial Application Field) The present invention provides a method for measuring an object using a two-dimensional visual device, even though the visual coordinate plane and the object coordinate plane are parallel. In reality, there are cases where the coordinate plane is tilted, and in this case, an error occurs in the measurement of the object, so the present invention relates to a method for calibrating the coordinate system in such a case.

(Prior art) As the prior art related to the present invention, (1) Japanese Patent Application Laid-Open No. 60-25
No. 2914 and (2) Japanese Unexamined Patent Publication No. 60151712.

What is shown in (1) above is related to the conversion method of visual sensor coordinate information and robot reference coordinate system information, and there are at least three
It teaches the coordinate value of a point in the visual sensor coordinate system and its coordinate value in the robot reference coordinate system, and converts the position information in the visual sensor coordinate system to the position information in the robot robot's reference coordinate system.

Furthermore, what is shown in (2) is related to the robot visual coordinate system calibration method, and the three-dimensional attachment that occurs when fixing the fixed visual field to the robot coordinate system is determined by moving the robot. .

(Problems to be Solved by the Invention) However, in both of the above (1) and (2), since the coordinates are converted using a robot, the processing mechanism is complicated, and errors may occur. There is a problem that processing time is unicost.

In particular, the conversion method shown in (1) above does not take into account mounting errors of the camera (visual) or the inclination of the workpiece installation plane, so there is a problem that even if the coordinates are converted, errors will occur and handling errors will occur. There is.

The present invention uses two-dimensional vision to provide a technology for easily calibrating the coordinate system when the coordinate plane is tilted in the actual process and a position error occurs when measuring an object. This is a major issue.

[Structure of the invention] (Means for solving the problem) This seems to be a proposal for a technical means to solve the problem. That is, it consists of a visual sensor that is the TV left camera, an image processing device that processes the position information of the visual sensor, and a monitor TV that displays the visual sensor information, and the two-dimensional position information of the jig placed within the field of view of the TV camera. is loaded into the memory of the image processing device, the center of gravity of the four holes of the jig is determined, and the position information of the visual sensor coordinate system and the coordinate system of the object is converted based on the distance between the centers of gravity. be.

(Function) Figure 3: A conceptual diagram showing the relationship between visual coordinates and object coordinates.
x, y, z are visual coordinates, which are the coordinates of the left TV camera in FIG. 2, and χ, y, z are object coordinates, which represent the coordinates of the workpiece in FIG.

The origin O of x, y, z is 13, the origin O" of object coordinates is 12
The displacement 0-0' of the origin of visual coordinates and object coordinates is indicated by hector 11.

If the object coordinates have an inclination or are not parallel to the visual coordinates, the object coordinates are inclined to the visual coordinates, and the object coordinates are
rotation angles α and β about the axes, respectively.

In this case, α is 14, β is 15,
T is indicated by 16. When the rotation angles α, β, and T are 0, both coordinates indicate a relationship of parallel movement.

Here, the visual coordinates X, Y, Z and the object coordinate χ.

The relationship between y and z is expressed by the following formula.

Here, Mα1Mβ, MT is a matrix representing the rotational displacement between both coordinates, and the rotation angle is α2β as shown in FIG.

If it is T, it becomes the following.

(COSβ Osinβ) Mβ=Omi 1 01 'sinβ Q cosβJ Since the above equation 0 is a general equation, the equation is simplified in order to apply it to two-dimensional vision. In other words, in two-dimensional vision, Z-
0, and since there is no error at all even if it is assumed that T is O, it is assumed that r=o, and that the object and the visual origin are also coincident. By the above simplification, the formula 0 becomes the following.

■If we expand and rearrange 2, we get the following formula S.

X-1cos3~zsinα Y-χsinαsinβ±y cosα-ZSjnα5
inB f■Z-1sInβcos a + y s
in αA-z cos αcosβ J2=0 From the above, if the unknowns α and β are obtained in advance, any visual coordinate value can be calibrated to the object coordinate.

(Example) Examples will be described below.

Figure 1 is a diagram showing the configuration of the visual measurement system, in which the visual mount on the workbench 8 is attached to the visual (TV left camera 2);
A measurement object (work) 4 placed within the visual field is measured, and the data is sent to an image processing device 6 via a cable 5, where measurements and the like are performed. Further, the processing contents are displayed on the TV monitor.

Figure 2 is an explanatory diagram of objects and visual coordinates, 9 is the visual coordinate,
Reference numeral 10 indicates object coordinates, and indicates a case where the visual attachment is normal (no rotation) with respect to the object coordinates.

4 and 5 explain the calibration procedure of this embodiment.

FIG. 4 shows a calibration jig, which has four holes with the same diameter, the distance between the opposing holes being constant, and the holes being perpendicular to each other. Calibrate using this jig as follows.

(1) Place the jig in the visual processing area shown in FIG.

(2) Store the jig image visually (TV camera) in the memory of the image processing device.

(3) At this time - For boat fishing, the holes are stored as being tilted as shown in 23 in Figure 5, but the centers of gravity of the two holes facing the χ and y axes of the visual coordinates are parallel to each other. put.

In this case, the center of gravity is calculated by an image processing device using a known algorithm, and displayed for each hole on a TV monitor, and the jig is moved so that the χ or y coordinates of the opposing holes are the same.

(4) Calculate the distances parallel to the χ and y axes.

(5) Assign a value to the 0 expression.

(6) ΔX (distance between jigs known in advance) - (χ center of gravity of hole 4 - χ center of gravity of hole 3) Find β from cos β.

(7) Similarly, from ■2, calculate α using (6).

By the above calculation, β of the rotation angle of the coordinate axes can be calculated. Also, this allows any coordinate value of the visual coordinate system to be calibrated with respect to the object coordinate value.

In addition, if the height of the object placed at Z-O in equation (2) is known in advance, it can be easily calibrated using the height information in exactly the same manner as in (2-0).

(Effects) The present invention has the following effects. That is, the visual sensor (T
V camera) When installed, workpieces are often tilted or placed at an angle with respect to the camera on actual lines, but the present invention can calibrate a two-dimensional coordinate system with a relatively simple configuration and method. Therefore, the object can be accurately measured and chuck errors can be eliminated even when gripping is performed by a robot system.

[Brief explanation of drawings]

Figure 1 is a visual measurement system configuration diagram, Figure 2 is an explanatory diagram of object coordinates and visual coordinates, Figure 3 is a general relationship diagram of object and visual coordinates, Figure 4 is a diagram of a coordinate calibration jig, and Figure 5 is a diagram of the coordinate calibration jig. The figure is a visual image explanatory diagram. 2... Visual sensor, 4... Work, 6... Image processing device, 7... Monitor television, 17... Jig.

Claims (1)

[Claims] It consists of a visual sensor consisting of a TV camera, an image processing device that processes the positional information of the visual sensor, and a monitor television that displays the visual sensor information, and processes the two-dimensional positional information of the jig placed within the field of view of the TV camera. A two-dimensional coordinate system calibration method that is loaded into the memory of the device, determines the center of gravity of the four holes of the jig, and converts position information between the visual sensor coordinate system and the coordinate system of the object based on the distance between the centers of gravity.