TWI701123B - Automated calibration system and method for workpiece coordinate frame of a robot - Google Patents

Abstract

An automated calibration system for workpiece coordinate frame of a robot is disclosed. The system mainly includes a physical image sensor which has a first image central axis, and a controller. The controller controls the physical image sensor adapted on a robot to move together to set up a virtual image sensor which also has a second image central axis, and these two image central axes have an intersection point. The controller controls the robot to repeatedly move back and forth a characteristic point on the workpiece between these two image central axes until the characteristic point coincides the intersection point. The controller records a calibration point including the coordinates of the joints of the robot, then the controller moves another characteristic point and repeats the above steps to generate several other calibration points. According to the calibration points, the controller calculates the relative coordinates of a virtual tool center point and the workpiece to the robot.

Description

Automatic correction system and method of robot workpiece coordinate system

The invention relates to a coordinate system automatic correction system, in particular to an automatic robot workpiece coordinate system correction system. The invention also relates to a robot workpiece coordinate system automatic correction method implemented on the robot workpiece coordinate system automatic correction system.

With the advancement of science and technology, the application of robots in various industries has become more and more extensive. Generally speaking, robots have joint-type mechanical arms with multiple joints, and a tool or cutter, such as a welding tool or a drill, is set at the end. Tools, etc., and generate robot actions through manual teaching to achieve automated operation applications.

Before the robot performs operations, the position of the Tool Center Point (TCP) needs to be accurately calibrated in advance, so that the controller of the robot can make the tool run on the correct path according to the tool center point.

However, as the working path of the robot becomes more and more complex, the accuracy of the working path is affected by the accuracy of the robot, and the accuracy of the relative relationship between the workpiece coordinate system and the robot directly affects the accuracy of the robot's action, so the accuracy of the workpiece coordinate system becomes the robot An important indicator for precise operation.

At present, when the robot executes automation applications, it is first necessary to confirm the relative positional relationship between the position of the workpiece and the robot. However, due to the accuracy of the positioning device, or the manufacturing tolerance of the workpiece, the positional relationship will have errors. Therefore, the robot needs to go through Only by correcting the position of the workpiece can accurate coordinate values be obtained.

The traditional method of workpiece position correction requires manual teaching, moving the tool center to coincide with several designated points on the workpiece, and recording the coordinates to complete the correction of the workpiece position.

If it is manually corrected, the robot needs to be taught to make the tool center point coincide with multiple specified points on the workpiece coordinate system to complete the correction. However, this method is not only affected by the operator’s experience, but the operation process is also prone to poor robot absolute accuracy. Good or careless operation, causing the center point of the tool to collide with the workpiece and damage.

If the automatic method is used for calibration, the existing method has the following disadvantages: (1) Install an image sensor on the outside of the robot to correct the coordinates of the workpiece by coincident with the designated point or measure the actual distance of the target point; this method may not be able to obtain the designated point image due to different tools, hidden by the robot body structure or objects ； (2) Install the image sensor and first confirm the relative position of the image sensor and the robot with measuring equipment, or control the center of the tool to touch the workpiece, but the workpiece may be damaged and there may be operator errors; (3) Use the CAD file to obtain the relative distance between the robot and the calibration instrument or the workpiece, but the operation process of this method is time-consuming and the workpiece needs to have feature points for actual measurement to reduce the error.

The applicant filed a Republic of China invention patent application on September 29, 106 of the Republic of China, the application number is 106133775, and the title of the invention is "Robot Tool Center Point Calibration System and Method", and then on August 11, 108 of the Republic of China The announcement was approved, the announcement number is I668541, and the aforementioned approved case is referred to as the "previous case" below.

However, based on the R&D spirit of striving for excellence, the applicant will make innovations and improvements in the following directions: (1) Prevent the image sensor from being shielded by the robot body structure or tools and unable to obtain the image of the specified point; (2) Reduce the cost of system architecture.

Therefore, there is obviously a need for a calibration process that does not require actual contact with the workpiece, has no collision problems, is not covered by the body structure or tools, does not need to calibrate the relationship between the image sensor position and the robot before operation, and does not require additional installation on the workpiece. The image sensor, and the workpiece position correction can be completed through a one-time calibration process, which can effectively improve the various shortcomings of the known art "Robot workpiece coordinate system and method for automatic correction".

In one embodiment, the present invention provides a robot workpiece coordinate system automatic calibration system, including: a physical image sensor having a first image center axis, the physical image sensor is arranged on the robot flange surface; and a controller, Control the rotation of the physical image sensor and the robot to construct a virtual image sensor, which has a second image central axis, and the second image central axis and the first image central axis have an intersection; wherein, the controller controls the robot to make a workpiece The feature point repeatedly moves between the central axis of the first image and the central axis of the second image until it coincides with the intersection point, records a calibration point containing a plurality of joint coordinates of the robot, repeats the above steps to generate a plurality of calibration points, according to these The calibration point calculates the coordinates of the center point of the virtual tool and the coordinates of the workpiece.

In another embodiment, the present invention provides an automatic calibration method for a robot workpiece coordinate system, including: providing a physical image sensor having a first image center axis, and the physical image sensor is arranged on the flange surface of the robot Provide a controller to control the rotation of the physical image sensor and the robot to construct a virtual image sensor, which has a second image central axis, and the second image central axis and the first image central axis have an intersection point; control the robot to make its workpiece The feature point repeatedly moves between the central axis of the first image and the central axis of the second image until it coincides with the intersection point. Record the correction point containing the coordinates of the multiple joints of the robot; move to the next feature point of the workpiece, and repeat the above steps A plurality of correction points are generated; the coordinates of the center point of the virtual tool and the relative coordinates of the workpiece are calculated according to the correction points.

Hereinafter, each embodiment of the content of the present invention will be described in detail, and the drawings will be used as examples. In addition to these detailed descriptions, the present invention can also be widely implemented in other embodiments, and easy substitutions, modifications, and equivalent changes of any of the embodiments are included in the scope of the present invention and will be covered by the following patents. Prevail. In the description of the specification, in order to enable the reader to have a more complete understanding of the present invention, many specific details are provided; however, the present invention may still be implemented under the premise that some or all of these specific details are omitted. In addition, well-known steps or elements are not described in details to avoid unnecessary limitation of the present invention. The same or similar elements in the drawings will be represented by the same or similar symbols. It should be noted that the drawings are for illustrative purposes only, and do not represent the actual size or quantity of the components unless otherwise specified.

As shown in the system architecture of Figure 1, the robot workpiece coordinate system 1 is connected to the robot R for implementation. The calibration system 1 may include a physical image sensor 11, a virtual image sensor 12, and a controller 13. The controller 13 may be The controller of the robot R or another independent computer device is not limited by the present invention. The physical image sensor 11 may be a substantial camera or other similar image capturing device. The robot R is also called a robotic arm, which includes a body M, which includes a plurality of joints J1~J6; the virtual image sensor 12 is simulated by the physical image sensor 11 and does not exist physically; the robot For example, the workpiece coordinate system automatic correction system 1 can be used to calculate the coordinates of a virtual tool center point TCP and the coordinates of a workpiece W, and this virtual tool center point is later used for reference when installing a physical tool on the end of the robot.

It must be emphasized that the system and method provided by the present invention do not need to install physical tools during the calibration process, but use several characteristic points or designated points on the workpiece W for calibration. Many workpieces can be selected on the workpiece W. The position of the feature point WPi is not limited, such as the center of a circle or the intersection of a line and a plane, etc., as long as it is located within the intersection range of the visual field of the physical image sensor 11 and the virtual image sensor 12, such as the workpiece feature point shown in FIG. 1 WPi is located at the intersection of the three surfaces, of course it can also be a point at other locations.

The physical image sensor 11 has a first image center axis A. The physical image sensor 11 is disposed on the flange surface F at the end of the robot R. The flange surface F has a coordinate system (x _f -y _f -z _f ) , The field of view of the physical image sensor 11 covers the Z axis zf at the center of the flange surface F, and the Z axis zf is perpendicular to the horizontal plane formed by the X axis xf and the Y axis yf.

2A shows the vector orientation captured by the physical image sensor 11, and FIG. 2B shows the relationship between the orientation and the coordinate system (x _1C -y _1C- z _1C ) of the physical image sensor 11; FIGS. 2A and 2B As shown, before performing workpiece coordinate correction, it is necessary to calculate the conversion between the coordinate system of the robot R (x _R -y _R -z _R ) and the coordinate system of the physical image sensor 11 (x _1C -y _1C -z _1C ) relationship.

First, after the mobile robot R makes any designated point on the workpiece W enter any position within the field of view of the physical image sensor 11, the designated point is taken as the origin O (not shown) of an image coordinate system. This image coordinate The system refers to the coordinate system (x _1C -y _1C -z _1C ) formed by the image captured by the physical image sensor 11.

方向任意長度得到實體影像感測器11的投影座標點

，並假設該點空間向量為

。 After the origin O coincides with another designated point, the mobile robot R follows the coordinate system of the robot R

Obtain the projection coordinate point of the physical image sensor 11 at any length in the direction

, And assume that the point space vector is

.

方向移動任意長度得到實體影像感測器11的投影座標點

，並假設該點空間向量為

。 Similarly, after the origin O coincides with the designated point, follow the coordinate system of the robot R

Move any length in the direction to obtain the projection coordinate point of the physical image sensor 11

, And assume that the point space vector is

.

方向移動任意長度得到實體影像感測器11的投影座標點

，並假設該點空間向量為

。 Similarly, after the origin O coincides with the designated point, follow the coordinate system of the robot R

Move any length in the direction to obtain the projection coordinate point of the physical image sensor 11

, And assume that the point space vector is

.

_•

=0………………………………………………………… (1)

_•

=0………………………………………………………… (2)

_•

=0………………………………………………………… (3) 並求解得到常數向量

。 Using the vertical characteristics of the coordinate system, the following simultaneous equations are obtained:

_•

=0…………………………………………………… (1)

_•

=0…………………………………………………… (2)

_•

=0………………………………………………………… (3) and solve the constant vector

.

與第二組解

相差一負號，

，因此可利用影像中工件上任意二指定點的距離長短變化，判別沿著機器人R之座標系(x _R-y _R-z _R)移動時，此指定點為朝向或遠離實體影像感測器11之方向移動，進而判別正確的分支解(Branch Solution)。 There are two sets of solutions for this equation, the first set of solutions

Solution with the second set

A minus sign,

, So the distance between any two designated points on the workpiece in the image can be used to determine when moving along the coordinate system of the robot R (x _R -y _R -z _R ), this designated point is toward or away from the physical image sensor Move in the direction of 11 to determine the correct branch solution (Branch Solution).

……………… (4) 其中，

為實體影像感測器11相對於法蘭面F座標系之座標，

為法蘭面F相對於機器人R的基座座標系(xb-yb-zb)之座標。 Therefore, the coordinates of the physical image sensor 11 relative to the F coordinate system of the flange surface:

……………… (4) Among them,

Is the coordinate of the physical image sensor 11 relative to the F coordinate system of the flange surface,

It is the coordinate of the flange surface F relative to the base coordinate system of the robot R (xb-yb-zb).

……………………………………(5) 其中，

為沿著實體影像感測器11之座標系(x _1C-y _1C-z _1C)之移動量，

為沿著機器人R之座標系(x _R-y _R-z _R)之移動量。 At the same time, the conversion relationship between the robot R coordinate system and the physical image sensor 11 coordinate system can be obtained:

……………………………………(5) Among them,

Is the amount of movement along the coordinate system (x _1C -y _1C -z _1C ) of the physical image sensor 11,

Is the amount of movement along the coordinate system of the robot R (x _R -y _R -z _R ).

產生第二視角，以作為虛擬影像感測器12，如圖1所示，而虛擬影像感測器12實質上是由實體影像感測器11所模擬而成的。 When the conversion relationship between the physical image sensor 11 and the robot R coordinate system is established, the physical image sensor 11 can be rotated by an angle relative to any axis of the flange surface F according to any point in the field of view

The second viewing angle is generated to serve as the virtual image sensor 12, as shown in FIG. 1, and the virtual image sensor 12 is essentially simulated by the physical image sensor 11.

Please refer to FIGS. 3A and 3B. The method for obtaining the coordinates of the center point TCP of the virtual tool and constructing the virtual image sensor 12 is described as follows:

Step (a): Move the robot R to a position with any feature point WPi in the field of view of the physical image sensor 11, and assume that the coordinate of any other point D in the field of view relative to the flange surface F is D=[ 0, 0, D _Z ], as shown in Figure 3A;

軸）為軸心分別旋轉兩個角度以產生兩點座標C ₂, C ₃，並以C ₁、C ₂、C ₃三點座標所產生之圓弧計算出一圓心位置D ₁，如圖3B所示； Step (b): Use the physical image sensor 11 to obtain the coordinate of the feature point WPi relative to the image coordinate system as C ₁ , and then the physical image sensor 11 uses its Z axis (

Axis) is the axis center by rotating two angles to generate two-point coordinates C ₂ , C ₃ , and calculate a circle center position D ₁ with the arc generated by the three-point coordinates C ₁ , C ₂ , C ₃ , as shown in Figure 3B Shown

至一工具影像中心點I ₁向量

，並將相對於該向量

轉換為相對於法蘭面F之向量

；此工具影像中心點I ₁即是工具之座標軸Z軸z _tool與第一影像中心軸A之交點。 Step (c): Calculate the center position

To a tool image center point I ₁ vector

And will be relative to this vector

Converted to a vector relative to the flange face F

; The center point I _{1 of} the tool image is the intersection point between the coordinate axis Z of the _tool and the center axis A of the first image.

，其中L()為距離函數，並回到步驟(a)。直到圓心位置D ₁與該影像中心點I ₁之座標重合後，即可取得虛擬工具中心點座標I。可根據每次向量

的變化，調整函數L()之常數；以及 Step (d): Correct the coordinate of point D to

, Where L() is the distance function, and return to step (a). Until the coordinates of the circle center position D ₁ and the image center point I ₁ coincide, the virtual tool center point coordinates I can be obtained. According to each vector

The change of, adjust the constant of the function L(); and

產生第二視角，以作為虛擬影像感測器12。 Step (e): Rotate the physical image sensor 11 by an angle on any axis of the flange F

The second angle of view is generated to serve as the virtual image sensor 12.

Please refer to Figure 4, in a visual servoing manner, move the physical image sensor 11 to make a specified point or feature point on the workpiece fall on the axis of view. From the captured image information, it can be obtained that the specified point is relative to The coordinates of the image sensor coordinate system are used to calculate the servo movement direction to complete the robot movement control. The steps are as follows:

Step (a1): Obtain the coordinates (x_c1, y_c1) of the specified point relative to the physical image sensor 11 from the image information captured by the physical image sensor 11;

轉換為機械人R座標系之方向

，如圖4所示； Step (b1): Change the direction of movement

Converted to the direction of the robot R coordinate system

,As shown in Figure 4;

Step (c1): Make the robot R move in the above direction until the designated point reaches the coordinate axis of the physical image sensor 11 and stops;

Step (d1): If the designated point is not located at the center point I ₁ of the physical image sensor 11, go back to step (a1), and if the designated point is located at the axis intersection point, the visual servoing control procedure is completed.

Please refer to FIG. 5, the physical image sensor 11 has a first image central axis A; the virtual image sensor 12 has a second image central axis B, the first image central axis A and the second image central axis B are not mutually exclusive Parallel has an intersection I, and the physical image sensor 11 and the virtual image sensor 12 have an image overlap area IA, so that the physical image sensor 11 and the virtual image sensor 12 can form a 2.5D machine vision; The virtual image sensor 12 is an essentially non-existent camera or other similar device, but is simulated by the physical image sensor 11.

The controller 13 controls the robot R to rotate together with the physical image sensor 11 or the virtual image sensor 12 so that any feature point WPi on the workpiece W is repositioned between the first image central axis A and the second image central axis B The cover movement is shown in Fig. 1; in a preferred embodiment, the controller 13 can be various computer devices. The controller 13 records a correction point at that time when the feature point WPi of the workpiece coincides with the intersection point I of the first image central axis A and the second image central axis B, and then changes the posture of the robot R to record the next correction point. In this way, a plurality of correction points of the robot R under different postures are recorded; finally, the controller 13 respectively calculates the coordinates of the virtual tool center point TCP and the workpiece W relative to the robot R according to the correction points; wherein, each correction point may include The coordinates of all joints J1~J6, and the coordinates of each joint can be the rotation angle of each joint relative to a preset starting point, that is, a correction point can represent a set of joint values; for example, if the joint angle θ (Joint angle ) Represents the coordinates of the joints, so the coordinates of most joints J1~J6 can be expressed as θ _{J 1} , θ _{J 2} , θ _{J 3} , θ _{J 4} , θ _{J 5} and θ _{J 6} ; so a correction point can be expressed as (θ _{J 1} , θ _{J 2} , θ _{J 3} , θ _{J 4} , θ _{J 5} , θ _{J 6} ).

It can be seen from FIG. 5 that the controller 13 first controls the robot R to move the workpiece feature point WPi to any position in the image overlapping area IA of the first image center axis A and the second image center axis B; then, the controller sequentially 13 Control the robot R to move the workpiece feature point WPi to the center axis A of the first image to point T1, as shown by the path PH1; then, the controller 13 controls the robot R to make the workpiece feature point WPi start from point T1 to the center of the second image Axis B moves to point T2, as shown by path PH2; similarly, the controller 13 controls the robot R to move the workpiece feature point WPi from point T2 to the central axis A of the first image to point T3, as shown by path PH3, and then Control the robot R to move the workpiece feature point WPi from point T3 to the second image center axis B to point T4, as shown by the path PH4; finally, the controller 13 controls the robot R to make the workpiece feature point WPi from point T4 to the first The image center axis A gradually moves to the intersection point I, and the first correction point CP1 is recorded when the workpiece feature point WPi coincides with the intersection point I; in this embodiment, the distance between the workpiece feature point WPi and the first image center axis A, and If the distance between the workpiece feature point WPi and the second image center axis B is less than a threshold value, it can be regarded as the workpiece feature point WPi coincides with the intersection point I; generally speaking, this threshold value can be set as the pixel of the workpiece feature point WPi 50%; that is, when the workpiece feature point WPi coincides with the first image central axis A and the second image central axis B within 50% of its pixels, it can be regarded as the workpiece feature point WPi coincides with the intersection point I; from the above, the controller 13 Control the robot R to repeatedly move the workpiece feature point WPi between the first image center axis A and the second image center axis B to obtain the first correction point CP1.

以修正機器人R的方位角，藉此改變機器人R的姿態；此時，機器人R之方位角可表示為

)，其中

為機器人R原來的方位角；其中R _x表示偏航角(Yaw angel)；R _y表示螺距角或俯仰角(Pitch angel)；R _z表示滾轉角(Roll angel)。若修正的方位角超過機器人R的運動範圍或超出交疊區域IA時，控制器13可透過亂數產生器重新再產生方位角增量。 Next, the controller 13 determines whether the number of correction points is greater than or equal to a preset value; in this embodiment, the number of correction points needs to be greater than or equal to 3; if the number of correction points is less than 3, the controller 13 can transmit the The number generator generates Euler Angle increments

To correct the azimuth angle of the robot R, thereby changing the posture of the robot R; at this time, the azimuth angle of the robot R can be expressed as

),among them

Is the original azimuth angle of the robot R; where R _x represents the yaw angle (Yaw angel); R _y represents the pitch angle or pitch angle (Pitch angel); R _z represents the roll angle (Roll angel). If the corrected azimuth angle exceeds the motion range of the robot R or exceeds the overlap area IA, the controller 13 can regenerate the azimuth angle increment through the random number generator.

Then, after obtaining the new azimuth angle and the next feature point WPi, the controller 13 controls the robot R to make the virtual tool center point TCP repeatedly move between the first image center axis A and the second image center axis B, and in the virtual The second calibration point CP2 is recorded when the tool center point TCP coincides with the second image center axis B.

Then, the controller 13 determines whether the number of calibration points is greater than or equal to 3; if the controller 13 determines that the number of calibration points is less than 3, the controller 13 repeats the above steps to obtain and record the third calibration point CP3, until the controller 13 determines the calibration The number of points is greater than or equal to 3.

According to the above description, the designated points or feature points used in the calibration process of the present invention may include at least 3 designated points that are known to be relative to the workpiece coordinate system, for example, the origin of the workpiece coordinate system, or any point on the X axis of the workpiece coordinate system. One point, any point on the XY plane of the workpiece coordinate system, etc. Move the robot R so that the i-th designated point of the workpiece coordinate system (ie the i-th feature point WPi of the workpiece W) is within the overlapping visual range of the physical image sensor 11 and the virtual image sensor 12, and repeat the above movement to overlap Step: Make i greater than a predetermined number to complete the designated point calibration information collection process, and calculate the coordinates of the virtual tool center and the workpiece based on the calibration points.

As shown in Fig. 1, the controller 13 can calculate the coordinates of the center point TCP of the virtual tool according to the calibration points CP1~CP3; the coordinates of the calibration points CP1~CP3 can be determined by the connecting rod parameters of the robot R (Denavit-Hartenberg Parameters) , The coordinates of the joints J1~J6 and the information of the coordinate system (x _f -y _f -z _f ) of the virtual tool center point TCP relative to the flange surface F are obtained; among them, the link parameters can include the link offset d(Link offset), joint angle θ (Joint angle), link length a (Link length), and link twist α (Link twist).

The coordinates of the central point TCP of the virtual tool can be calculated by the following formula (6): T _1i T ₂ =P………………………………………………(6)

……………….…………(7) Among them, the matrix T _1i is to convert the coordinates of the i-th calibration point from the coordinate system of the base (x _b -y _b -z _b ) to the coordinate system of the flange surface F (x _f -y _f -z _f ) Matrix T ₂ is the coordinate system of the virtual tool center point TCP relative to the flange surface F, and matrix P is the coordinate system of the calibration point relative to the base in space (x _b- y _b -z _b ); each correction point can be used to obtain three linear equations through equation (6), so 3n equations can be obtained using n correction points, and then the virtual tool center can be obtained by the pseudo-inverse matrix (Pseudo-inverse matrix) Point the coordinates of TCP; from equation (6), equation (7) can be derived:

……………….…………(7)

…………………….…..(8)

………………..…………………………(9) Among them, (e _11i , e _21i , e _31i ) represents the direction of the vector of the i-th correction point on the x _f axis relative to the coordinate system of the base (x _b -y _b -z _b ); (e _12i , e _22i , e _32i ) represents the direction of the vector of the i-th correction point on the y _f axis relative to the coordinate system of the base (x _b -y _b -z _b ); (e _13i , e _23i , e _33i ) represents the i-th The direction of the vector of the correction point on the z _f axis with respect to the coordinate system (x _b -y _b -z _b ) of the base; from equation (7), equations (8) and (9) can be derived:

…………………….…..(8)

………………..…………………………(9)

，

為

之轉置矩陣(Transpose matrix)，

為

之反矩陣(Inverse matrix)。 among them,

,

for

Transpose matrix (Transpose matrix),

for

The inverse matrix.

移項後得出式(9)，取得虛擬工具中心點TCP相對於法蘭面F之座標系之座標(T _x, T _y, T _z)及虛擬工具中心點TCP相對於機器人R座標系(x _R-y _R-z _R)之座標(P _x, P _y, P _z)，並完成虛擬工具中心點TCP之座標(T _x, T _y, T _z)之校正以及計算出工件W之座標。 If the number of correction points is sufficient, use the known matrix T _1i corresponding to the i-th correction point to substitute each element in the matrix into equation (8) and the matrix

After shifting the term, formula (9) is obtained, and the coordinates of the virtual tool center point TCP relative to the coordinate system of the flange surface F (T _x , T _y , T _z ) and the virtual tool center point TCP relative to the robot R coordinate system (x _R -y _R -z _R ) coordinates (P _x , P _y , P _z ), and complete the correction of the coordinates (T _x , T _y , T _z ) of the virtual tool center point TCP and calculate the coordinates of the workpiece W.

It can be seen from the above that in this embodiment, the robot workpiece coordinate system automatic correction system 1 can automatically correct the coordinates of the robot relative to the processed workpiece through visual servoing, and can achieve extremely high correction accuracy, so it can effectively reduce Labor cost and time cost; In addition, the robot workpiece coordinate system automatic correction system 1 can accurately correct the robot's coordinate relative to the workpiece through a calibration procedure. Therefore, the robot workpiece coordinate system automatic correction system 1 can indeed effectively improve the shortcomings of the prior art.

Please refer to FIG. 6, the method of obtaining the conversion relationship between the coordinate system (x _R -y _R -z _R ) of the robot R and the coordinate system (x _1C -y _1C -z _1C ) of the physical image sensor 11 may include The following steps:

Step S51: Control the robot R to move any designated point within the field of view of the physical image sensor 11 from any position in the image overlap area IA along the coordinate system (x _R -y _R -z _R ) of the robot R The horizontal axis x _R moves a distance L _R , and the physical image sensor 11 obtains the first projection coordinate P′ _x1 .

Step S52: Control the robot R to move the designated point from the position in the image overlap area IA along the longitudinal axis y _R of the coordinate system (x _R -y _R -z _R ) of the robot R by a distance L _R , and The physical image sensor 11 obtains the second projection coordinate P′ _y1 .

Step S53: Control the robot R to move the designated point from the position in the image overlap area IA along the vertical axis z _R of the coordinate system (x _R -y _R -z _R ) of the robot R by a distance L _R , and The physical image sensor 11 obtains the third projection coordinate P′ _z1 .

、第二空間向量

及第三空間向量

。 Step S54: providing a first projection corresponding to the coordinate P _'x1, a second projection coordinates P' _Y1 and the third projection coordinates P 'of the first space vector _z1

, Second space vector

And third space vector

.

、第二空間向量

及第三空間向量

之間的垂直關係計算第一空間向量

、第二空間向量

及第三空間向量

。 Step S55: According to the first space vector

, Second space vector

And third space vector

The vertical relationship between the calculation of the first space vector

, Second space vector

And third space vector

.

、第二空間向量

及第三空間向量

計算機器人R之座標系(x _R-y _R-z _R)與實體影像感測器11之座標系(x _1C-y _1C-z _1C)之轉換關係，如上述式(4)。 Step S56: According to the first space vector

, Second space vector

And third space vector

The conversion relationship between the coordinate system (x _R -y _R -z _R ) of the robot R and the coordinate system (x _1C -y _1C -z _1C ) of the physical image sensor 11 is calculated, as shown in the above formula (4).

Please refer to FIG. 7, the method of obtaining the conversion relationship between the coordinate system (x _R -y _R -z _R ) of the robot R and the coordinate system (x _2C -y _2C -z _2C ) of the virtual image sensor 12 may include The following steps:

Step S61: Control the robot R to move any designated point in the field of view of the virtual image sensor 12 from any position in the image overlapping area IA along the horizontal axis of the coordinate system (x _R -y _R -z _R ) of the robot R x _R moves a distance L _R , and the virtual image sensor 12 obtains the first projection coordinate P′ _x2 .

Step S62: Control the robot R to move the designated point from the position in the image overlap area IA along the longitudinal axis y _R of the coordinate system (x _R -y _R -z _R ) of the robot R by a distance L _R , and The virtual image sensor 12 obtains the second projection coordinate P′ _y2 .

Step S63: Control the robot R to move the designated point from the position in the image overlap area IA along the vertical axis z _R of the coordinate system (x _R -y _R -z _R ) of the robot R by a distance L _R , and The virtual image sensor 12 obtains the third projection coordinate P′ _z2 .

、第二空間向量

及第三空間向量

。 Step S64: Provide a first space vector corresponding to the first projection coordinate _P'x2 , the second projection coordinate _P'y2 and the third projection coordinate _P'z2

, Second space vector

And third space vector

.

、第二空間向量

及第三空間向量

之間的垂直關係計算第一空間向量

、第二空間向量

及第三空間向量

。 Step S65: According to the first space vector

, Second space vector

And third space vector

The vertical relationship between the calculation of the first space vector

, Second space vector

And third space vector

.

、第二空間向量

及第三空間向量

計算機器人R之座標系(x _R-y _R-z _R)與虛擬影像感測器12之座標系(x _2C-y _2C-z _2C)之轉換關係，如上述式(8)。 Step S66: According to the first space vector

, Second space vector

And third space vector

The conversion relationship between the coordinate system (x _R -y _R -z _R ) of the robot R and the coordinate system (x _2C -y _2C -z _2C ) of the virtual image sensor 12 is calculated, as shown in the above formula (8).

Please refer to Figure 8, the method adopted by the robot workpiece coordinate system automatic correction system 1 includes the following steps:

Step S71: Provide a physical image sensor 11, which has a first image center axis A.

Step S72: Provide a virtual image sensor 12, which has a second image central axis B, and the second image central axis B and the first image central axis A have an intersection point I.

Step S73: Control the robot R to move the feature point WPi of the workpiece repeatedly between the central axis A of the first image and the central axis B of the second image.

Step S74: When the workpiece feature point WPi coincides with the intersection point I, record the correction point including the coordinates of the multiple joints J1~J6 of the robot R.

Step S75: Select the next workpiece feature point WPi, and repeat the above steps to generate a plurality of correction points.

Step S76: Calculate the coordinates of the center point TCP of the virtual tool and the workpiece coordinates according to the correction points.

Please refer to FIG. 9 which illustrates the flow chart of the automatic calibration method of the robot workpiece coordinate system of the present invention in more detail:

Step S81: assuming i=1, set the i-th feature point WPi of the workpiece W.

Step S82: The controller 13 moves the robot R so that the feature point WPi of the workpiece is located in the common field of view of the physical image sensor 11 and the virtual image sensor 12.

Step S83: The controller 13 first moves the workpiece feature point WPi to the first image central axis A, and then to the second image central axis B, and moves repeatedly until the workpiece feature point WPi is respectively connected to the first image central axis A and the first image central axis A The central axis B of the two images coincides with the intersection point I.

Step S84: The controller 13 determines whether the distance error between the intersection point I of the first image center axis A and the second image center axis B and the feature point WPi of the workpiece is less than a threshold; if yes, proceed to step S85; if not, proceed Step S841.

)以修正機器人R的方位角，並回到步驟S83。 Step S841: The controller 13 generates the azimuth angle increment through the random number generator (

) To correct the azimuth angle of the robot R, and return to step S83.

Step S85: The controller 13 records the first set of joint values of the workpiece feature point WPi, that is, the first correction point.

Step S86: The controller 13 determines whether the number of joint values in the first group is greater than or equal to four. If yes, proceed to step S87; if not, proceed to step S841; the number can be other numbers, and the present invention is not limited.

Step S87: The controller 13 determines whether the number of workpiece feature points WPi is greater than or equal to the specified number; if yes, proceed to step S88; if not, proceed to step S871.

Step S871: Let i=i+1.

Step S88: Obtain the coordinates of the virtual tool center point TCP and the workpiece W relative to the robot R in the tool center correction method; wherein the tool center correction method has been disclosed in the previous proposal, and the present invention will not be described again.

In summary, in the system and method for automatic calibration of robot workpiece coordinates provided by the present invention, the calibration process can be divided into four parts: (1) Mounting a physical image sensor on the flange surface of the robot end; (2) Establish the conversion relationship between the robot flange surface coordinate system and the physical image sensor coordinate system, so that the motion information obtained by the image is converted into the motion information of the robot; (3) Construct the virtual image sensor and The position of the center point of the virtual tool generates 2.5D machine vision; (4) Through the image servo method, the robot is controlled to make the specified point on the workpiece coincide with the intersection of the two image central axes. This can be accomplished by one of the following methods: (41) Control the robot as an example More than four different postures make the center point of the virtual tool coincide with the origin of the workpiece coordinate, and then use any posture to make the center point of the virtual tool coincide with any point on the X axis of the workpiece coordinate system, any point on the XY plane, and record its coordinates, or (42) For example, the control robot uses more than four different postures to make the center point of the virtual tool coincide with more than four known coordinate points on the workpiece relative to the workpiece coordinate system, and record their coordinates.

The present invention only needs to install a physical image sensor on the flange surface of the robot end, and perform calibration through the feature points of the workpiece. The calibration process does not need to actually contact the workpiece, and there is no collision problem. The workpiece can be completed through a single calibration process. The position correction effectively improves the accuracy of the correction.

Although the present invention has been disclosed in the above embodiments, it is not intended to limit the present invention. Anyone with ordinary knowledge in the technical field can make some changes and modifications without departing from the spirit and scope of the present invention. The scope of protection of the present invention shall be determined by the scope of the attached patent application.

、

、

、

、

、

:空間向量 T1~T4:移動點 TCP:虛擬工具中心點 W:工件 WPi:工件特徵點 x₁~x₃、y₁~y₃、z₁~z₃、x_C1、y_C1、x_C2、y_C2、C₁、C₂、C₃:座標 x_C、y_C、z_C、x_R、y_R、z_R、x_b、y_b、z_b、x_f、y_f、z_{f 、}x_tool、z_tool:座標軸θv:角度1: Robot workpiece coordinate system automatic calibration system 11: Physical image sensor 12: Virtual image sensor 13: Controller A: The first image central axis B: The second image central axis CP1~CP3: Calibration point R: Robot F: Flange surface I: Intersection point IA: Overlapping area J1~J6: Joint M: Body _P'x1 , _P'y1 , _P'z1 , _P'x2 , _P'y2 , _P'z2 : Projection coordinates S51~S56 , S61~S66, S71~S76, S81~S87: step flow S _C1 , S _C2 , PH1~PH4: distance

,

,

,

,

,

: Space vector T1~T4: Moving point TCP: Virtual tool center point W: Workpiece WPi: Workpiece feature point x ₁ ~ x ₃ , y ₁ ~ y ₃ , z ₁ ~ z ₃ , x _C1 , y _C1 , x _C2 , y _C2 , C ₁ , C ₂ , C ₃ : coordinates x _C , y _C , z _C , x _R , y _R , z _R , x _b , y _b , z _b , x _f , y _f , z _{f ,} x _tool , z _tool : coordinate axis θv : angle

1 is a schematic diagram of the system architecture of the first embodiment of the robot workpiece coordinate system automatic correction system of the present invention. 2A~2B are schematic diagrams of the conversion relationship of the first embodiment of the robot workpiece coordinate system automatic correction system of the present invention. 3A to 3B are schematic diagrams of the construction steps of the first embodiment of the robot workpiece coordinate system automatic correction system of the present invention. 4 is a schematic diagram of the visual servoing of the first embodiment of the robot workpiece coordinate system automatic correction system of the present invention. Fig. 5 is a schematic diagram of the moving and overlapping steps of the first embodiment of the robot workpiece coordinate system automatic correction system of the present invention. Fig. 6 is the first flow chart of the first embodiment of the robot workpiece coordinate system automatic correction system of the present invention. Fig. 7 is a second flowchart of the first embodiment of the robot workpiece coordinate system automatic correction system of the present invention. Fig. 8 is a third flowchart of the first embodiment of the robot workpiece coordinate system automatic correction system of the present invention. Fig. 9 is a fourth flowchart of the first embodiment of the robot workpiece coordinate system automatic correction system of the present invention.

no

1: Robot tool center point correction system

11: Physical image sensor

12: Virtual image sensor

13: Controller

F: Flange surface

J1~J6: Joint

M: body

R: Robot

TCP: Virtual Tool Central Point

W: Workpiece

WPi: Workpiece feature points

x _b , y _b , z _b , x _f , y _f , z _f , x _tool , z _tool : coordinate axis

θv : angle

Claims (20)

A robot workpiece coordinate system automatic correction system, suitable for connecting a robot, the system includes: A physical image sensor having a first image center axis, the physical image sensor being arranged on a flange surface of the end of the robot; and A controller is connected to the robot and the physical image sensor, and controls the rotation of the physical image sensor and the robot to construct a virtual image sensor. The virtual image sensor has a second image center axis. The central axis of the second image and the central axis of the first image have an intersection; Wherein, the controller controls the robot to make a feature point on a workpiece repeatedly move between the center axis of the first image and the center axis of the second image until it coincides with the intersection point, and records the number of joints including the robot One of the coordinate correction points, and then by moving to the next feature point, repeat the above movement to overlap the intersection point to generate a plurality of other correction points, according to the correction points to calculate a virtual tool center point and the workpiece relative to the robot The coordinates. The robot workpiece coordinate system as described in the first item of the scope of patent application, wherein the controller converts the coordinate system of the robot to one of the coordinate system of the physical image sensor and the virtual image sensor Relationship, and a plurality of images of the physical image sensor and the virtual image sensor to control the movement of the robot. As described in the first item of the scope of patent application, the robot workpiece coordinate system is an automatic correction system, wherein each of the correction points includes the rotation angle of the joints relative to a preset starting point. The robot workpiece coordinate system described in item 3 of the scope of patent application is an automatic calibration system, wherein the controller calculates the coordinates of the center point of the virtual tool according to the calibration points and a link parameter of the robot. For example, the robot workpiece coordinate system described in item 1 of the scope of patent application is an automatic correction system, wherein the number of the correction points is greater than or equal to a preset value. For example, the robot workpiece coordinate system described in item 5 of the scope of patent application is an automatic correction system, wherein when the number of correction points is less than the preset value, the controller generates an azimuth angle increment through a random number generator to correct One of the azimuths of the robot. For example, the robot workpiece coordinate system described in item 6 of the scope of patent application is an automatic correction system, wherein the controller, after correcting the azimuth angle of the robot, controls the robot so that another feature point is on the central axis of the first image and the first image The two image center axes are moved repeatedly to obtain another calibration point until the number of the calibration points is greater than or equal to the preset value. For example, the robot workpiece coordinate system described in item 1 of the scope of patent application is an automatic correction system, wherein when the characteristic point coincides with the intersection of the central axis of the first image and the central axis of the second image, the characteristic point and the first image The distance between the central axis and the distance between the feature point of the workpiece and the central axis of the second image are respectively smaller than a threshold value. The robot workpiece coordinate system described in item 1 of the scope of patent application is an automatic correction system, wherein the coordinate of the virtual tool center point is relative to the coordinate of a base of the robot, or relative to the flange surface. A method for automatically correcting the coordinate system of a robot workpiece is suitable for a robot connected to a controller. The method includes the following steps: (i) Provide a physical image sensor forming an image coordinate system and having a first image center axis, and the physical image sensor is arranged on a flange surface of the end of the robot; (ii) The controller controls the rotation of the physical image sensor and the robot to construct a virtual image sensor, the virtual image sensor having a second image center axis, the second image center axis and the The central axis of the first image has an intersection; (iii) The controller controls the robot to make a feature point on a workpiece repeatedly move between the central axis of the first image and the central axis of the second image until it coincides with the intersection point, and records a plurality of points including the robot One of the calibration points of the joint coordinates; (iv) The controller controls the robot to move to the next feature point, repeating the above movement to coincide with the intersection point to generate a plurality of other correction points; and (v) Calculate the center point of a virtual tool and the coordinates of the workpiece relative to the robot according to the correction points. For example, the automatic calibration method of the robot workpiece coordinate system described in item 10 of the scope of patent application further includes a method of constructing the virtual image sensor. The steps are: (a) Move the robot so that the characteristic point falls within the field of view of the physical image sensor, and obtain the coordinates of an arbitrary point in the field of view relative to the flange surface; (b) Obtain the first point of one of the coordinate systems of the arbitrary point relative to the physical image sensor, and then use the central axis of the first image as the center to rotate two angles to generate the second point and the third point, And calculate a center position with the first point, the second point, and the third point; (c) Calculate the vector from the center position of the circle to the center point of a tool image, and convert the vector relative to the image coordinate system into a vector relative to the flange surface. The center point of the tool image is a coordinate axis of a tool and The intersection of the central axis of the first image; (d) Correct the coordinates of the arbitrary point, and return to step (a) until the center position of the circle coincides with the coordinates of the center point of the tool image, then obtain a tool center point coordinate; and (e) Rotate the physical image sensor by an angle according to any axis of the flange surface to serve as the virtual image sensor. The automatic calibration method of the robot workpiece coordinate system as described in item 10 of the scope of patent application, wherein the step (iii) further includes the following steps: Provide a conversion relationship between the coordinate system of the robot and the coordinate system of the physical image sensor and the virtual image sensor; and According to the conversion relationship, the multiple images of the physical image sensor and the virtual image sensor, the robot is controlled to move. The method for automatically calibrating the coordinate system of a robot workpiece as described in item 12 of the scope of patent application, wherein the step of providing the conversion relationship between the coordinate system of the robot and the coordinate system of the physical image sensor and the virtual image sensor It also contains: Control the robot to move the feature point at any position in the overlapping area of an image of the physical image sensor and the virtual image sensor for a distance along the horizontal axis of the coordinate system of the robot, and the physical The image sensor and the virtual image sensor obtain a first projection coordinate; Control the robot to move the feature point from the position of the image overlapping area by the distance along the longitudinal axis of the robot's coordinate system, and obtain a second from the physical image sensor and the virtual image sensor Projection coordinates; and Control the robot to move the feature point from the position of the image overlapping area by the distance along the vertical axis of the robot's coordinate system, and obtain a third from the physical image sensor and the virtual image sensor Projection coordinates. The method for automatically calibrating the coordinate system of a robot workpiece as described in item 13 of the scope of patent application, wherein the step of providing the conversion relationship between the coordinate system of the robot and the coordinate system of the physical image sensor and the virtual image sensor It also contains: Providing a first space vector, a second space vector, and a third space vector corresponding to the first projection coordinate, the second projection coordinate, and the third projection coordinate respectively; Calculating the first space vector, the second space vector, and the third space vector according to the vertical relationship between the first space vector, the second space vector, and the third space vector; and According to the first space vector, the second space vector, and the third space vector, the conversion relationship between the coordinate system of the robot and the coordinate system of the physical image sensor and the virtual image sensor is calculated. As described in item 10 of the scope of patent application, the coordinates of each of the joints are the rotation angles of each of the joints relative to a predetermined starting point. The automatic calibration method of the robot workpiece coordinate system described in item 10 of the scope of patent application, wherein the step (v) further includes the following steps: According to the calibration points and a link parameter of the robot, the coordinates of the center point of the virtual tool and the coordinates of the workpiece are calculated. For example, in the automatic calibration method of the robot workpiece coordinate system described in item 10 of the scope of patent application, the number of the calibration points is greater than or equal to a preset value. The automatic calibration method of the robot workpiece coordinate system as described in item 10 of the scope of patent application, wherein the step (iii) further includes the following steps: The distance between the characteristic point and the central axis of the first image and the distance between the characteristic point and the central axis of the second image are respectively smaller than a threshold value. The automatic calibration method of the robot workpiece coordinate system as described in item 10 of the scope of patent application, wherein the step (iv) further includes the following steps: When the number of the correction points is less than a preset value, an azimuth angle increment is generated through the random number generator to correct an azimuth angle of the robot. As described in item 19 of the scope of patent application, the automatic correction method of the robot workpiece coordinate system includes the following steps after correcting the azimuth angle of the robot: The robot is controlled to repeatedly move another characteristic point between the central axis of the first image and the central axis of the second image to obtain another correction point until the number of the correction points is greater than or equal to the preset value.

Images