KR920006741B1 - Noncontact type tire contour shape measurement system - Google Patents

Abstract

This non-contact measuring device of configuration of any object comprises a senser section (40) which is composed of a slit light generator (41) and camera (42); a computer which sends to CRT (60) a three dimentional coordinate value of the object and calibrates from the senser; a robot driver who controls the robotic hand by multiple axes in accordance with the control signal from the computer. The robotic driver includes the X-Z axis serbo-controller which is controlled by the data signal and address of the computer, α,β axis controllers, and actual performance is done by X-Z axis direction setting serbo-motor and α,β

Description

Profile measuring device for non-contact tires

1 is a block diagram of an apparatus of the present invention.

2 is an explanatory view of a non-contact tire profile contour measurement by the sensor unit of the present invention.

3 is an explanatory diagram of a slit light generating device of the sensor unit of the present invention.

4 is an explanatory diagram for calculating a maximum curvature point for extracting feature points from a calibration block.

5 is an explanatory diagram of various parameters obtained from a calibration block.

6 is a configuration diagram of the multi-axis robot hand of the present invention and a related arrangement of the sensor unit, the measurement object and the calibration block.

7 is a block diagram of the robot hand drive unit of the present invention.

8 is an explanatory diagram of a change in measurement accuracy according to the distance between the sensor unit camera and the measurement object of the present invention.

9 is an explanatory diagram for removing the blind spots when measuring the position movement in the object to be measured according to the present invention.

10 is a calibration flowchart according to the present invention.

11 is a measurement flowchart according to the present invention.

* Explanation of symbols for main parts of the drawings

10 computer 11: robot control unit

12: I / O port 13: image processing unit

14: CPU 20: robot driver

21-23: X-Y axis servo controller 24,25: α, β axis controller

30: robot hand 40: sensor

41: slit light generator 42: camera

45: support 60: CRT

90: object to be measured 100: calibration block

The present invention relates to an apparatus for accurately measuring the shape or the overall shape of a specific part of an object without error, and in particular, continuous measurement with a single camera on any measurement object such as a tire using a non-contact slit light shape detection sensor The present invention relates to a cross-sectional contour shape measuring apparatus of a non-contact tire which enables to increase the accuracy and speed of three-dimensional measurement of an object by eliminating rectangular generation during measurement.

In general, when measuring a specific object by using a contact shape detection sensor which is a common measurement method of the object, if the measurement object is a relatively solid solid, the measurement error does not occur significantly, but the measurement object is a semi-solid state Or when the deformation comes to the original shape due to the contact of the sensor such as a tire or the like, a considerable measurement error is caused.

In particular, when measuring the outer sectional contour of a tire composed of both side walls and treads, the contact sensor has a defect in the measurement itself, for example, a conventional non-contact measurement. Even if the equipment is used, it is impossible to ignore the hassles and measurement errors that need to be compared several times by using various instruments to find spatial three-dimensional coordinate values for each part.

SUMMARY OF THE INVENTION An object of the present invention is to provide a three-dimensional simultaneous shape measurement apparatus that can significantly improve the measurement efficiency and measurement accuracy of an object while eliminating all the problems in the existing object shape measurement apparatus as described above.

A feature of the present invention is to use a slit light and a camera to generate a two-dimensional coordinate value of an object to be measured and a calibration block, and a two-dimensional measurement coordinate value obtained from the sensor part to correspond to the two-dimensional coordinate value of the calibration block. A computer that outputs a three-dimensional shape to a CRT or the like by substituting and calculating a spatial three-dimensional coordinate value, a robot hand for multi-axis transformation of the measurement position of the sensor unit, and a robot driver for controlling the robot hand according to the control signal of the computer. It can be expressed as including.

Details of these features of the present invention are as described in claim 2 below.

Hereinafter, an embodiment of the present invention will be described with reference to the accompanying drawings.

As shown in FIG. 1, the robot control unit 11 and the CPU of the computer 10 including the robot control unit 11, the I / O port 12, the image processing unit 13, and the CPU 14 are shown. (14) is configured to send and receive data with the robot driver 20, the robot hand 30 driven by the robot driver 20 is configured to control the position of the non-contact shape measurement sensor unit 40, The shape measurement signal of the non-contact sensor unit 40 is configured to be output to the CRT 60, the printer 70, the plotter 80, etc. via the image processor 13.

In addition, the I / O port 12 is configured to input a shape measurement execution signal or data by the keyboard 50.

2 shows the configuration of the sensor unit 40, the shape measurement, the video signal, and the transmission system by the sensor unit 40. As shown in FIG.

Here, the slit light generator 41 and the camera 42 are fixed on the support 45, and the measurement object 90 or slit light (plane light) is irradiated from the slit light generator 41 or the like. The camera 42 scans the contour of the shadow that appears according to the shape of the calibration block 100 (shown in FIG. 5) and outputs it to the CRT 60 or the like.

As shown in FIG. 3, the slit light generating device causes the laser beam generated by the laser to be converted into slit light while passing through the cylindrical lens.

6 shows the robot hand 30 in detail.

As can be seen here, the robot hand 30 installs a vertical pole 33 on the movable plate 32 moving in the XY axis direction on the base 31, and one side of the upper end of the vertical pole 33. The other end of the horizontal bar 35 having the first motor holder 38 at one end using a holder 34 is fixed.

The first motor holder 38 is provided with the first motor 36 in the vertical direction, so that the sensor unit 40 including the second motor 37 coupled to the second motor holder 39 is simultaneously in the left and right directions. It is configured to be movable, and the second motor 37 is installed in the horizontal direction, the sensor unit 40 coupled to the support 45 is configured to be movable in the vertical direction.

In addition, in FIG. 6, the sensor unit 40 moves the position of the measurement target object 90 and the calibration block 100 on the calibration table 101 that illustrate the tire as an example, so that the slit light can be irradiated and photographed. It is arranged to be.

7 shows the servo motors 311-313 and encoders 314-316 installed in the base 31 of the robot hand 30, and the first and second stepping motors 36 and 37, respectively. 20 to be controlled by the XZ axis servo controller 21-23 and the α, β axis controllers 24 and 25, and the XZ axis servo controller 21-23 and the α, β axis controller 24 And 25 are controlled to be controlled by the address and data of the CPU 14 in the computer 10.

Referring to the operation and effects of the embodiment of the present invention configured as described above are as follows.

Since the sensor unit 40 of the non-contact cross-sectional contour measuring device is composed of a set in which the slit light generating device 41 and the camera 42 are fixed in a constant shape as shown in FIG. 2, the measurement of the measurement target object 90 is performed. When the slit light is shined on the site, a contour shape is formed at the site where the measurement object 90 and the slit light meet. By reading this with the camera 42, two-dimensional image data as shown on the CRT 60 can be obtained. .

By analyzing this two-dimensional image data, three-dimensional position data of each point of the slit light irradiation site is obtained, and shape measurement of the measurement target object 90 such as a tire is performed.

The three-dimensional spatial coordinates are obtained from the calibration block 100 by correspondingly obtaining calibration parameter values between two-dimensional coordinates in the camera system.

The camera egg is a generic term for a device using one or a plurality of cameras or a device using an auxiliary device such as a plane light irradiation device. In order to calibrate a camera system, it is necessary to know three-dimensional spatial coordinate values for at least some points and coordinate values in the camera system corresponding to each point.

When the image of the edge or the step of the calibration block 100 is inputted to the camera 42 and analyzed by the computer 10, the end and the edge of the staircase are perpendicular to each other. The coordinates of the camera system for each point on the coordinates can be found as feature points. In addition, since the dimensions from the reference plane to the steps or the edges are already known, the three-dimensional coordinate values for each step and the edge can be known if only the three-dimensional coordinate values of the reference plane are known.

Looking at the calibration of the camera system for the interpretation of the image data, as shown in Fig. 2, when a point P (x, y, z) in space forms on the image plane of the camera, the coordinate in the image system is P '. (UV), when P 'is expressed in Homogenous Coordinate (u, v, h),

Is related to.

P and P '

The relation is formed and from equation (3) and (4)

Substituting equations (1) and (2) into (5) and (6) respectively

Substituting equation (7) into equations (8) and (9), respectively,

Equations (10) and (11) can be used in the following three cases.

Case 1) Coordinate value on the image plane: When the transformation matrix T and x, y, z are known in advance, two coordinates (U, V) are unknown so the coordinate value of the image plane can be known.

Case 2) Measurement: When the transformation matrixes T, U, and V are known in advance, the solution cannot be obtained because there are three unknowns (x, y, z) and two equations. Therefore, the solution can be solved by adding more equations by Stereo Vision or Slit Ray Projection.

Case 3) Calibration: When we know the position (x ₁ , y ₁ , z ₁ ) of several points in 3D space and know the image coordinates (U ₁ , V ₁ ) of each point, we can find the transformation matrix T. have.

Since equations (10) and (11) have two equations and there are l2 unknowns, the spatial and image coordinate values (x ₁ , y _l , z ₁ , U ₁ , V ₁ ) for six points are need. (i = 1, ~, 6)

Equations (10) and (11) are homogeneous equations, so if we assume C ₄₃ = 1, there are 11 unknowns, so that unknowns can be obtained with 5.5 or more data.

If you express this as a matrix

In short

If T is obtained by Pseudo-Inverse in Equation (13),

Looking at the calibration by planar tube projection, one plane is determined by three points and the equation is

Is represented by

It is expressed as

Since the equations are insufficient to determine the unknown by the camera system itself during measurement, three-dimensional measurements are made by simultaneously using the above-described camera system correction and slit light correction by adding plane light correction, and obtaining the correction parameters accordingly.

From the camera-related equations of equations (3) and (4) and the slit beam equation of equation (16)

From equation (17)

In relation to equations (17) and (18)

From equations (18) and (19)

If we know U, V, x, y, z, we have 12 unknowns (if M ₃₄ = 1, 11) and we have 3 equations, so we need to know the spatial and image coordinates for at least four points.

If you express this as a matrix,

(25)

Expressed in, unknown m is by the shadow inverse

It is calculated as

Once m is determined, when the coordinates U and V of the points appearing in the image coordinate system are known, the coordinates in the three-dimensional space of the corresponding point can be known therefrom.

That is, when M ₃₄ = 1 from equations (22), (23) and (24)

In order to obtain m _ij for the three-dimensional measurement using the sensor unit 40 consisting of the slit light generator 41 and the camera 42 from equation (25), data for at least four points are required during the calibration process. Able to know. Each of these four points must know in advance the three-dimensional coordinates of the system in space, so that the slit beams of the steps of the calibration block 100 of FIGS. In this case, the image of the slit light as shown in FIG. 5 is read.

At this time, x, y, z value of each known feature point in spatial coordinate system (x _l , y _l , z ₁ , x ₂ , y ₂ , z ₂ , x ₃ , y ₃ , z ₃ , x ₄ , y ₄ , z ₄ ) and the coordinate values U and V (U ₁ , V ₁ , U ₂ , V ₂ , U ₃ , V ₃ , U ₄ , V ₄ ) in the image system (inside the sensor sub-system) are substituted into the equation (26). Then, the sensor unit is calibrated by obtaining the m value as shown in the equation (27).

To do this, the edges (in the CRT) of the imaging system are distinguished, and the two-dimensional coordinate values (U, V values) in the imaging system of each point are identified, and then the three-dimensional coordinate values (x, y) in space , z value) and the camera coordinate system value must be matched.

The three-dimensional coordinate values of the corner points in space can be known by the actual measurement in the actual coordinate system, and the corner determination for obtaining the value in the image coordinate system of the corner points uses a method of extracting the curvature maximum point.

The curvature C _ik at point P ₁ shown in FIG.

Y ', Y ", Y ₁ are also calculated in the same way.

From these the maximum curvature C _im

The point P ₁ whose maximum curvature is C _{im is} calculated from and becomes the maximum curvature if the following conditions are met.

I-j for all j ≤ m / 2

By the above-described curvature maximal point extraction method, a feature point (edge point) represented by each U and V of the video system shown in the CRT 60 is found.

The calibration of the sensor unit 40 is performed, and the conversion parameter M (Equation 18 m matrix) is obtained. The calibration is completed. Based on this, the spatial position of each point of the object to be measured is expressed in Equations 28, 29 and 30. Is measured by.

However, one of the items affecting the measurement accuracy of the system is the resolution of the camera 42, which is one of the components of the sensor unit 40. For example, an approximate length of a measured object processed by one screen is 50 m. / m and the scanning line of the camera 42 is about 500 lines, the resolution is limited to 0. lm / m. By doing this, the length of the part to be measured actually exceeds the length that can be grasped by the camera 42 at one time. As shown in FIG. You can increase the accuracy of the measurement by close-up. As described above, in the case where it is difficult to measure at once, the method of performing the measurement while connecting the sensor unit 40 appropriately and connecting each other is used as in FIG.

Referring to the measurement flowchart of FIG. 11, the money plate control position in the computer 10 and the robot driver 20 according to the movement position path specification of the money manipulator (Manipulator) input by the keyboard 50 or the like in FIG. The image is acquired by the slit light generating device 41 and the camera 40 of the sensor unit 40 by the generation and the movement of the money plate, that is, the movement of the robot hand 30.

By extracting the center line from such a two-dimensional image and applying the parameters of the sensor unit 40 to it, three-dimensional (shown as 3D in the drawings) information is extracted and stored, and then the measurement data is joined to output the data Or control the robot hand's movement again or terminate.

Prior to the measurement of such a measurement object, the same calibration as that of the flowchart of FIG. 10 is performed. During calibration, move the money plate to the calibration position and set the position of the calibration block.

When the calibration image is obtained at the calibration position of the money plater, the feature point is extracted from the contour center line, and the parameters of the sensor unit are extracted and stored. At this time, the 3D coordinates of the feature point are referred to. These parameters are evaluated and, if appropriate, terminated and if not, move the position of the money plate back to the calibration position.

On the other hand, the measurement as described above not only improves the measurement accuracy but also excludes the possibility of blind spots that may occur during measurement.

The exclusion of the blind spots is sufficiently explained by the fact that the measurement range of the measurement target object is overlapped when the camera 42 moves by the robot hand as shown in FIG.

Measurement by the movement method of the sensor part 40 is performed by attaching the sensor part 40 containing the camera 42 to the end of the robot hand 30 like FIG.

In this case, since the sensor part is moved and rotated by three axes of Cartesian coordinates of X, Y, and Z, and two axes of rotation of α, β, consideration should be given to this. For example, at any point on the tire, the measured value is (x ₁ , y _l , z ₁ ), where the amount of movement with respect to each axis (X, Y, Z) is (p, q, r) and the amount of rotation is α ₁ , β _The actual measurement (x ₂ , y ₂ , z ₂ ) taking into account the robot movement when

Where T = (p, q, r): movement of x, y, z axis

R is a matrix representing rotation, and as in FIG. 6, the amount of rotation on the α axis corresponds to the amount of rotation parallel to the Z axis, and the amount of rotation on the β axis depends on the amount of rotation on the α axis (90 or 180 degrees). Corresponds to the amount of rotation parallel to the axis. In other words

For convenience, the reference system of the measured coordinate values is based on the system based on the coordinate values obtained at the time of calibration. For example, if the tire tread is measured, α = 90, β = 90 ° (X-axis)

Referring to Figures 6 and 7 attached to the modification control process of the multi-axis robot hand 30 as described above are as follows.

Since each data and address are applied from the computer 10 to the XZ axis servo controllers 21-23 and the α and β axis controllers 24 and 25 in the robot driver 20 through the respective buses, the XY The axis servo controllers 21-23 control the respective servo motors 311-313 inside the base 31 in accordance with the data from the computer 10.

Accordingly, since the movable plate 32 and the vertical pole 33 move in the X, Y and Z axes, respectively, the first and second motor holders 38 and 39 and the support stand 45 are placed on the horizontal bar 35. The sensor unit 40 coupled and fixed through the scan moves while moving the measurement target object 90.

In addition, the encoders 314-316 added to the servo motors 311-313 encode rotation angles of the servo motors, and provide the encoded servo angles to the servo controllers. Is continuously controlled in such a way that the servo motor stops when the servo motor is rotated by a predetermined limit amount.

On the other hand, the α and β axis controllers 24 and 25 of the robot driver 20 are connected to the first stepping motor 36 and the second stepping motor through the harmonic drivers 361 and 371 (represented in HD in the drawing). 37, the control signal is supplied. Accordingly, the first motor 36 fixed to the holder 38 at one end of the horizontal bar 35 includes the second motor 37 to provide the sensor unit 40 with the control signal. The sensor unit 40 makes the measurement target object 90 beta by the second motor 37 which rotates around the α axis and is installed on the β axis perpendicular to the α axis of the first motor 36. Rotate around the axis and scan its outer shape.

As described above, the present invention is equipped with a camera and a slit light generator in a computer-controlled multi-axis robot hand, and scans the cross-sectional shape of the object to be measured over time with a single camera, and calculates the corresponding process in the calibration block. Since the object to be measured can be represented on the space 3 coordinates, a unique effect of measuring the cross-sectional outer shape of a tire, such as a side circle or a tread, can be quickly measured without error.

Claims (3)

A device for measuring and recognizing an outer shape of an object in a non-contact manner, comprising: a sensor unit 40 including a slit light generator 41 and a camera 42, and a calibration block from the calibration block by the sensor unit 40; A computer which calculates three-dimensional spatial coordinates from two-dimensional calibration coordinates and performs corresponding calculation on two-dimensional measured coordinates from the measurement object and outputs the three-dimensional optical coordinates of the measurement object to CR T (60). 10), the robot hand 30 for multi-axis conversion of the measurement position of the sensor unit 40, and the robot driver 20 for multi-axis control of the robot hand 30 in accordance with the control signal of the computer 10 Cross-sectional contour shape measuring device of a non-contact tire, characterized in that it comprises. 2. The sectional contour shape measuring device according to claim 1, wherein the slit light generating device (41) generates slit light by passing a laser beam through a cylindrical lens. The robot driver (20) according to claim 1, wherein the robot driver (20) includes an XZ-axis servo controller (21-23) and an α-β axis controller (24, 25) controlled in accordance with the address and data signal of the computer 10. As an output of each of the XZ axis servo controllers 21-23, an XZ axis direction servo motor 311-313 is used, and as an output of each of the α and β axis controllers 24 and 25, A cross-sectional contour shape measuring device for a non-contact tire, characterized in that the axial movement stepping motors (36, 37) are configured to be controlled.

Images