CN104260112A - Robot hand and eye locating method - Google Patents

Abstract

The invention belongs to the technical field of mechanical automation, in particular to a robot hand and eye locating method. The robot hand and eye locating method includes that pasting a calibration plate on a reference workpiece, building a visual coordinate system, calculating coordinates of all the points on the reference workpiece under the visual coordinate system, moving the reference workpiece and the calibration plate to the downside of the robot station, and obtaining the coordinates of all the points on the reference workpiece through converting between the visual coordinate system and a robot coordinate system. The robot hand and eye locating method is suitable for the robot assembly line work featured with large locating range and high locating precision.

Description

A kind of Robot Hand-eye localization method

Technical field

The invention belongs to technical field of mechanical automation, relate to a kind of localization method, be specifically related to a kind of Robot Hand-eye localization method.

Background technology

Along with the continuous progress of science and technology, industrial automaticity is more and more higher, and robot is enhancing productivity, and guarantees that the effect on the quality of production, attenuating production cost is most important.Robot application is in process of production as generally, and the position how robot operationally obtains workpiece is accurately very important.

When traditional way is mounting robot on streamline, the precision relying on robot to install and the put precision of workpiece on streamline remove control in the course of the work to the location of workpiece; The drawback of this way is apparent, and the installation accuracy of robot and the precision of putting of workpiece are all Artificial Control, and therefore, the problem inaccurate to Workpiece fixing operationally often appears in robot, cannot finish the work efficiently.

In order to address this problem, at present, develop a kind of robot and the artificial technology positioned workpiece of collocation vision system replacement, this technology has now become a kind of classical mode on streamline.

Robot and the vision system mode of employing are in the market: on working robot, install vision system, during work, directly by vision system, workpiece is positioned, then vision system direct control unit device people carries out work, but the little streamline of orientation range that is only applicable to of this mode uses.

Therefore, present urgent problem is exactly: how to realize the streamline large to scope and accurately locate.

Summary of the invention

In order to solve the problem in background technology, the present invention proposes a kind of Robot Hand-eye localization method being applicable to the robot pipelining that orientation range is large, positioning precision is high.

Concrete technical scheme of the present invention is:

1, a Robot Hand-eye localization method, is characterized in that: comprise the following steps:

1) visual coordinate system is set up:

1.1) choosing benchmark workpiece any two points is shooting point, and all the other points on benchmark workpiece are calculation level; If any two points is designated as B ₁point and B ₂;

1.2) adjust the position of benchmark workpiece on streamline, two shooting point lines are projected in image coordinate system parallel with the axis of ordinates of image coordinate system;

1.3) on benchmark workpiece, scaling board is sticked; Described scaling board is provided with dot matrix; In dot matrix, the line of every a line point is all parallel with the axis of abscissas of image coordinate system, and each row point line is parallel with axis of ordinates; Any point in dot matrix is set to initial point, in initial point horizontal direction line be a little axis of abscissas, on initial point vertical direction line be a little axis of ordinates, set up visual coordinate system (x ^v, y ^v);

2) coordinate of each point under visual coordinate system is calculated:

2.1) according to the ratio calculation actual physical size coefficient of the difference of the pixel value of the actual range of any two points in dot matrix on scaling board and any two points of shooting; The concrete calculating formula of actual physical size coefficient is:

K _x＝L/M _x，K _y＝L/M _y

Wherein, on scaling board, in dot matrix, the actual range of any two points is definite value, is set to L;

M _xfor the difference of the pixel value on any two points abscissa direction of shooting, M _yfor the difference of the pixel value on any two points ordinate direction of shooting;

K _xfor the actual physical size coefficient on abscissa direction, K _yfor the actual physical size coefficient on ordinate direction is;

2.2) according to the B of shooting ₁the pixel of point and the pixel value of visual coordinate system initial point, calculate B ₁the coordinate of point under visual coordinate system, concrete calculating formula is:

$\left\{\begin{array}{c}{x}_{B1}^x=k_x\·(x_{B1}^p-x_0^p)\\ y_{B1}^v=-k_y\·(y_{B1}^p-y_0^p)\end{array}\right.---\left(1\right)$

Wherein, B ₁the pixel value of point is b ₁the coordinate of point under visual coordinate system is the pixel value of visual coordinate system initial point is

2.3) according to the practical structures size of benchmark workpiece, namely on benchmark workpiece all the other points relative on the position of B1 point and benchmark workpiece, all the other are put relative to B ₁and B ₂point line angle, to calculate on benchmark workpiece the coordinate a little under visual coordinate system: setting tool body calculating formula is:

$\left\{\begin{array}{c}{X}^v=\mathrm{dx}\·\cos \mathrm{\θ}-\mathrm{dy}\·\sin \mathrm{\θ}+x_{B1}^v\\ Y^v=\mathrm{dx}\·\sin \mathrm{\θ}+\mathrm{dy}\·\cos \mathrm{\θ}+y_{B1}^v\end{array}\right.---\left(2\right)$

(dx, dy) is for all the other points on benchmark workpiece are relative to the position of B1 point;

θ is that the line of all the other points and B1 point on benchmark workpiece is relative to B ₁and B ₂point line angle;

3) under benchmark workpiece being moved to robot station together with scaling board;

4) visual coordinate system is converted to robot coordinate system:

Judge that the type of robot is robot for punching or assembly robot; If robot for punching, then perform step 4.1; If assembly robot, then perform step 4.2;

4.1) to obtain on benchmark workpiece the coordinate a little under robot for punching coordinate system:

4.1.1) by benchmark workpiece movable under the station of robot for punching, the coordinate value of initial point under robot coordinate system in acquisition point matrix using coordinate value as the translational movement under visual coordinate ties up to robot coordinate system;

4.1.2) deflection angle of computation vision coordinate system axis of ordinates and robot coordinate system's axis of ordinates and the deflection angle of visual coordinate system axis of abscissas and robot coordinate system's axis of abscissas;

4.1.2.1) at least two row point lines on dot matrix are got, get the angle of each row point line and robot coordinate system's axis of ordinates respectively, according to the algorithm of averaging of suing for peace, obtain the deflection angle of visual coordinate system axis of ordinates and robot coordinate system's axis of ordinates;

4.1.2.2) at least two row point lines on dot matrix are got, get the angle of every a line point line and robot coordinate system's axis of abscissas respectively, according to the algorithm of averaging of suing for peace, obtain the deflection angle of visual coordinate system axis of abscissas and robot coordinate system's axis of abscissas;

4.1.3) according to the physical length of scaling board dot matrix mid point line and the projection relation of robot coordinate system, the length ratio of line under acquisition point line physical length and robot coordinate system, is put;

4.1.4) visual coordinate is tied to the conversion formula of robot coordinate and is:

$\left\{\begin{array}{c}{x}^r=(X^v\·\cos {\mathrm{\θ}}_y-Y^v\·{\cos \mathrm{\θ}}_y)/\cos \left({\mathrm{\θ}}_y\right)\·L_x^{\′}/L_x+x_0^r\\ y^r=-(Y^v\·{\cos \mathrm{\θ}}_x+X^v\·{\cos \mathrm{\θ}}_x)/\cos \left({\mathrm{\θ}}_x\right)\·L_y^{\′}/L_y+y_0^r\end{array}\right.---\left(3\right)$

Wherein: for the coordinate value of initial point in dot matrix under robot coordinate system;

L _x, L _yfor the physical length of dot matrix mid point line;

L' _x, L' _yfor the length of dot matrix mid point line under robot coordinate system;

θ _xfor the deflection angle of visual coordinate system axis of ordinates and robot coordinate system's axis of ordinates;

θ _yfor the deflection angle of visual coordinate system axis of ordinates and robot coordinate system's axis of ordinates;

4.2) to obtain on benchmark workpiece the coordinate a little under assembly robot coordinate system:

4.2.1) by benchmark workpiece movable under the station of assembly robot, adjustment robot end installing plate is parallel with benchmark workpiece, records robot end's anglec of rotation now;

4.2.2) in acquisition point matrix, the coordinate value of initial point under robot coordinate system ties up to the translational movement under robot coordinate system as visual coordinate;

4.2.3) angle between some line and the axis of ordinates of robot coordinate system arranged arbitrarily in the angle between some line and the axis of abscissas of robot coordinate system and dot matrix of going arbitrarily in dot matrix is got respectively; The algorithm computation vision coordinate system of averaging according to suing for peace and the deflection angle of robot coordinate system;

4.2.4) obtaining the conversion formula that visual coordinate is tied to robot coordinate system is:

$\left\{\begin{array}{c}{x}_1^r=x^r+L\·\cos ({\mathrm{\θ}}_1+{\mathrm{\θ}}_L)-L\·\cos ({\mathrm{\θ}}_1+{\mathrm{\θ}}_L+{\mathrm{\θ}}_0)\\ y_1^r=y^r+L\·\sin ({\mathrm{\θ}}_1+{\mathrm{\θ}}_L)-L\·\cos ({\mathrm{\θ}}_1+{\mathrm{\θ}}_L+{\mathrm{\θ}}_0)\end{array}\right.---\left(4\right)$

Wherein: L is the distance of alignment point in robot rotation axis points to robot end's installing plate;

θ _lfor the line of alignment point in robot end's rotation axis points and robot end's installing plate and the angle of visual coordinate system abscissa;

θ ₁for the deflection angle of visual coordinate system and robot coordinate system;

θ ₀for robot end's anglec of rotation;

5) robot picks up workpiece according to the coordinate of the workpiece obtained under robot coordinate system.

Above-mentioned image coordinate system is the upper left corner of the image taken is initial point, and the transverse direction of shooting image is axis of abscissas, and the longitudinal direction of shooting image is the coordinate system that axis of ordinates is set up.

The invention has the advantages that:

1, the present invention takes the method for visual coordinate system converting machine people coordinate system, makes robot when carrying out pipelining on a large scale, accurate positioning.

2, the present invention analyzes the Robot Vision Calibration method of two kinds of techniques, is applicable to vision robot's application of the various uses such as arc-welding, carrying, piling, vanning, assembling.

Accompanying drawing explanation

Fig. 1 is the front view of benchmark workpiece;

Fig. 2 is the schematic diagram of dot matrix on scaling board;

Fig. 3 is the graph of a relation of robot for punching coordinate system and visual coordinate system;

Fig. 4 is the graph of a relation of assembly robot coordinate system and visual coordinate system.

Detailed description of the invention

The present invention proposes the Robot Hand-eye localization method that a kind of orientation range is large, positioning precision is high, below in conjunction with accompanying drawing, technical scheme of the present invention be described:

Due to camera camera site and robot operating position different, operated by robot after needing to transmit certain distance on streamline after vision system shooting location completes, before robot work, need the conversion carrying out visual coordinate system and robot coordinate system.

Step 1) set up visual coordinate system;

Step 1.1) as shown in Figure 1: choosing benchmark workpiece any two points is shooting point, and all the other points on benchmark workpiece are calculation level; If any two points is designated as B ₁point and B ₂point;

Step 1.2) position of benchmark workpiece on adjustment streamline, make two shooting point (B ₁point and B ₂point) line projects parallel with the axis of ordinates of image coordinate system in image coordinate system;

Wherein, image coordinate system is exactly the upper left corner of the image taken is initial point O, and the transverse direction of shooting image is the axis of abscissas X of image coordinate system, and the longitudinal direction of shooting image is the coordinate system (X, Y) of the axis of ordinates Y foundation of image coordinate system;

Step 1.3) on benchmark workpiece, stick scaling board (figure of scaling board is see Fig. 2); Scaling board is provided with dot matrix (OABC); Guarantee that the line of every a line point on scaling board is all parallel with the X-axis of image coordinate system, each row point line is parallel with Y-axis; Any point in dot matrix is set to initial point, in initial point horizontal direction line be a little axis of abscissas, on initial point vertical direction line be a little axis of ordinates, set up visual coordinate system (x ^v, y ^v);

As shown in Figure 2, using O point as visual coordinate system initial point, the line of O point and A point is axis of ordinates, and the line of O point and B point is axis of abscissas,

Step 2) calculate the coordinate of each point under visual coordinate system:

Step 2.1) according to the ratio calculation actual physical size coefficient of the difference of the pixel value of the actual range of any two points in dot matrix on scaling board and any two points of shooting; The concrete calculating formula of actual physical size coefficient is:

K _x＝L/M _x，K _y＝L/M _y

Wherein, on scaling board, in dot matrix, the actual range of any two points is definite value, is set to L;

M _xfor the difference of the pixel value on any two points abscissa direction of shooting, M _yfor the difference of the pixel value on any two points ordinate direction of shooting;

K _xfor the actual physical size coefficient on abscissa direction, K _yfor the actual physical size coefficient on ordinate direction;

Step 2.2) according to the B taken ₁the pixel of point and the pixel value of visual coordinate system initial point, calculate B ₁the coordinate of point under visual coordinate system, concrete calculating formula is:

$\left\{\begin{array}{c}{x}_{B1}^x=k_x\·(x_{B1}^p-x_0^p)\\ y_{B1}^v=-k_y\·(y_{B1}^p-y_0^p)\end{array}\right.---\left(1\right)$

Wherein, B ₁the pixel value of point is b ₁the coordinate of point under visual coordinate system is the pixel value of visual coordinate system initial point is

Step 2.3) according to the practical structures size of benchmark workpiece, namely on benchmark workpiece all the other points relative on the position of B1 point and benchmark workpiece, all the other are put relative to B ₁and B ₂point line angle, to calculate on benchmark workpiece the coordinate a little under visual coordinate system: setting tool body calculating formula is:

$\left\{\begin{array}{c}{X}^v=\mathrm{dx}\·\cos \mathrm{\θ}-\mathrm{dy}\·\sin \mathrm{\θ}+x_{B1}^v\\ Y^v=\mathrm{dx}\·\sin \mathrm{\θ}+\mathrm{dy}\·\cos \mathrm{\θ}+y_{B1}^v\end{array}\right.---\left(2\right)$

(dx, dy) is for all the other points on benchmark workpiece are relative to the position of B1 point;

θ is that the line of all the other points and B1 point on benchmark workpiece is relative to B ₁and B ₂point line angle;

Step 3) at completing steps 2) after, under benchmark workpiece is moved to robot station together with scaling board;

Step 4) visual coordinate system is converted to robot coordinate system:

There are two types in general robot: complete the robot for punching of the hole operation that screws and complete the assembly robot of installment work; Wherein, robot for punching only need provide coordinate, and assembly robot, except providing coordinate, also provides the anglec of rotation of robot end's installing plate.

Therefore, completing steps 4) when needing, need the type first judging robot, the type of robot is divided into two kinds of situations to carry out the conversion that visual coordinate is tied to robot coordinate system:

Judge that the type of robot is robot for punching or assembly robot; If robot for punching, then perform step 4.1); If assembly robot, then perform step 4.2);

Step 4.1) to follow these steps to obtain on benchmark workpiece the coordinate a little under robot for punching coordinate system:

Because punch machines robot end has installed cylinder and screw fixing machine.Cause the uneven situation of the plane of robot coordinate plane and benchmark workpiece like this;

Step 4.1.1) by benchmark workpiece movable under the station of robot for punching, obtaining step 1.3) in the coordinate value of dot matrix Central Plains point (as shown in Figure 2, i.e. O point) under robot coordinate system using coordinate value as the translational movement under visual coordinate ties up to robot coordinate system;

Step 4.1.2) computation vision coordinate system axis of ordinates and the deflection angle of robot coordinate system's axis of ordinates and the deflection angle of visual coordinate system axis of abscissas and robot coordinate system's axis of abscissas;

Step 4.1.2.1) get at least two row point lines on dot matrix, get the angle of each row point line and robot coordinate system's axis of ordinates respectively, according to the algorithm of averaging of suing for peace, obtain the deflection angle of visual coordinate system axis of ordinates and robot coordinate system's axis of ordinates;

Step 4.1.2.2) get at least two row point lines on dot matrix, get the angle of every a line point line and robot coordinate system's axis of abscissas respectively, according to the algorithm of averaging of suing for peace, obtain the deflection angle of visual coordinate system axis of abscissas and robot coordinate system's axis of abscissas;

As shown in Figure 3 and Figure 4: first, robot moves and obtains vision calibration plate A, B, C point coordinates,

A) calculate OA and BC two straight lines under robot coordinate system with the deflection angle mean value θ of y-axis _y(θ _yjust be to be partial on the left of y-axis, right side is negative);

B) OB and AC two straight lines under robot coordinate system with the deflection angle mean value θ of x-axis _x(θ _xjust be to be partial on the downside of x-axis, upside is negative);

Step 4.1.3) according to the physical length of scaling board dot matrix mid point line and the projection relation of robot coordinate system, under acquisition point line physical length and robot coordinate system, put the length ratio of line;

Specific practice is: owing to there is projection relation, and therefore projected length also exists a proportionality coefficient, according to the line segment length L' under robot coordinate system's axis of abscissas that OABC obtains _x, line segment length L' under axis of ordinates _ywith the physical length lateral length L of OABC _x, longitudinal length L _x; And calculate abscissa proportionality coefficient and ordinate proportionality coefficient respectively;

Step 4.1.4) completing steps 4.1.1) to step 4.1.3) after, the conversion formula that acquisition visual coordinate is tied to robot coordinate is:

$\left\{\begin{array}{c}{x}^r=(X^v\·\cos {\mathrm{\θ}}_y-Y^v\·{\cos \mathrm{\θ}}_y)/\cos \left({\mathrm{\θ}}_y\right)\·L_x^{\′}/L_x+x_0^r\\ y^r=-(Y^v\·{\cos \mathrm{\θ}}_x+X^v\·{\cos \mathrm{\θ}}_x)/\cos \left({\mathrm{\θ}}_x\right)\·L_y^{\′}/L_y+y_0^r\end{array}\right.---\left(3\right)$

Put together machines artificial four articulated robots, not only needs to provide coordinate value, also need to provide the end anglec of rotation during work;

Step 4.2) to follow these steps to obtain on benchmark workpiece the coordinate a little under assembly robot coordinate system:

Step 4.2.1) by benchmark workpiece movable under the station of assembly robot, adjustment robot end installing plate is parallel with benchmark workpiece, now accurately can be placed on benchmark workpiece by module to be assembled, records robot end anglec of rotation θ now ₀;

Step 4.2.2) obtaining step 1.3) in the coordinate value of dot matrix Central Plains point (as shown in Figure 2, i.e. O point) under robot coordinate system using coordinate value as the translational movement under visual coordinate ties up to robot coordinate system;

Step 4.2.3) to get in dot matrix the angle between some line and the axis of ordinates of robot coordinate system that angle between the some line of row arbitrarily and the axis of abscissas of robot coordinate system and scaling board arrange arbitrarily in dot matrix respectively; The algorithm computation vision coordinate system of averaging according to suing for peace and the deflection angle of robot coordinate system;

Specific practice is: shown in composition graphs 2 and Fig. 4; The installing hole put together machines on the installing plate of robot end is respectively to initial point O, A, B point in dot matrix, obtain the coordinate of three points in robot coordinate system, calculate initial point O and B under robot coordinate system and put the angle of line and X-axis, datum point O and A puts and Y-axis angle, calculating mean value θ ₁, be the anglec of rotation of robot coordinate system relative to visual coordinate system.

Step 4.2.4) completing steps 4.2.1) to step 4.2.3) after, the conversion formula that acquisition visual coordinate is tied to robot coordinate system is:

$\left\{\begin{array}{c}{x}_1^r=x^r+L\·\cos ({\mathrm{\θ}}_1+{\mathrm{\θ}}_L)-L\·\cos ({\mathrm{\θ}}_1+{\mathrm{\θ}}_L+{\mathrm{\θ}}_0)\\ y_1^r=y^r+L\·\sin ({\mathrm{\θ}}_1+{\mathrm{\θ}}_L)-L\·\cos ({\mathrm{\θ}}_1+{\mathrm{\θ}}_L+{\mathrm{\θ}}_0)\end{array}\right.---\left(4\right)$

Wherein, L is the line of alignment point on robot end's rotation axis points and installing plate on installing plate;

θ _lfor the line of alignment point on robot end's rotation axis points on installing plate and installing plate and the angle of visual coordinate system abscissa;

θ ₁for the deflection angle of visual coordinate system and robot coordinate system;

θ ₀for robot end's anglec of rotation.

Step 5) robot picks up workpiece according to the coordinate of the workpiece obtained under robot coordinate system.

Claims (2)

1. a Robot Hand-eye localization method, is characterized in that: comprise the following steps:

1) visual coordinate system is set up:

1.1) choosing benchmark workpiece any two points is shooting point, and all the other points on benchmark workpiece are calculation level; If any two points is designated as B ₁point and B ₂;

1.2) adjust the position of benchmark workpiece on streamline, two shooting point lines are projected in image coordinate system parallel with the axis of ordinates of image coordinate system;

1.3) on benchmark workpiece, scaling board is sticked; Described scaling board is provided with dot matrix; In dot matrix, the line of every a line point is all parallel with the axis of abscissas of image coordinate system, and each row point line is parallel with axis of ordinates; Any point in dot matrix is set to initial point, in initial point horizontal direction line be a little axis of abscissas, on initial point vertical direction line be a little axis of ordinates, set up visual coordinate system (x ^v, y ^v);

2) coordinate of each point under visual coordinate system is calculated:

2.1) according to the ratio calculation actual physical size coefficient of the difference of the pixel value of the actual range of any two points in dot matrix on scaling board and any two points of shooting; The concrete calculating formula of actual physical size coefficient is:

K _x＝L/M _x，K _y＝L/M _y

Wherein, on scaling board, in dot matrix, the actual range of any two points is definite value, is set to L;

M _xfor the difference of the pixel value on any two points abscissa direction of shooting, M _yfor the difference of the pixel value on any two points ordinate direction of shooting;

K _xfor the actual physical size coefficient on abscissa direction, K _yfor the actual physical size coefficient on ordinate direction is;

2.2) according to the B of shooting ₁the pixel of point and the pixel value of visual coordinate system initial point, calculate B ₁the coordinate of point under visual coordinate system, concrete calculating formula is:

$\left\{\begin{array}{c}{x}_{B1}^x=k_x\·(x_{B1}^p-x_0^p)\\ y_{B1}^v=-k_y\·(y_{B1}^p-y_0^p)\end{array}\right.---\left(1\right)$

Wherein, B ₁the pixel value of point is b ₁the coordinate of point under visual coordinate system is the pixel value of visual coordinate system initial point is

2.3) according to the practical structures size of benchmark workpiece, namely on benchmark workpiece all the other points relative on the position of B1 point and benchmark workpiece, all the other are put relative to B ₁and B ₂point line angle, to calculate on benchmark workpiece the coordinate a little under visual coordinate system: setting tool body calculating formula is:

$\left\{\begin{array}{c}{X}^v=\mathrm{dx}\·\cos \mathrm{\θ}-\mathrm{dy}\·\sin \mathrm{\θ}+x_{B1}^v\\ Y^v=\mathrm{dx}\·\sin \mathrm{\θ}+\mathrm{dy}\·\cos \mathrm{\θ}+y_{B1}^v\end{array}\right.---\left(2\right)$

(dx, dy) is for all the other points on benchmark workpiece are relative to the position of B1 point;

θ is that the line of all the other points and B1 point on benchmark workpiece is relative to B ₁and B ₂point line angle;

3) under benchmark workpiece being moved to robot station together with scaling board;

4) visual coordinate system is converted to robot coordinate system:

Judge that the type of robot is robot for punching or assembly robot; If robot for punching, then perform step 4.1; If assembly robot, then perform step 4.2;

4.1) to obtain on benchmark workpiece the coordinate a little under robot for punching coordinate system:

4.1.1) by benchmark workpiece movable under the station of robot for punching, the coordinate value of initial point under robot coordinate system in acquisition point matrix using coordinate value as the translational movement under visual coordinate ties up to robot coordinate system;

4.1.2) deflection angle of computation vision coordinate system axis of ordinates and robot coordinate system's axis of ordinates and the deflection angle of visual coordinate system axis of abscissas and robot coordinate system's axis of abscissas;

4.1.2.1) at least two row point lines on dot matrix are got, get the angle of each row point line and robot coordinate system's axis of ordinates respectively, according to the algorithm of averaging of suing for peace, obtain the deflection angle of visual coordinate system axis of ordinates and robot coordinate system's axis of ordinates;

4.1.2.2) at least two row point lines on dot matrix are got, get the angle of every a line point line and robot coordinate system's axis of abscissas respectively, according to the algorithm of averaging of suing for peace, obtain the deflection angle of visual coordinate system axis of abscissas and robot coordinate system's axis of abscissas;

4.1.3) according to the physical length of scaling board dot matrix mid point line and the projection relation of robot coordinate system, the length ratio of line under acquisition point line physical length and robot coordinate system, is put;

4.1.4) visual coordinate is tied to the conversion formula of robot coordinate and is:

$\left\{\begin{array}{c}{x}^r=(X^v\·\cos {\mathrm{\θ}}_y-Y^v\·{\cos \mathrm{\θ}}_y)/\cos \left({\mathrm{\θ}}_y\right)\·L_x^{\′}/L_x+x_0^r\\ y^r=-(Y^v\·{\cos \mathrm{\θ}}_x+X^v\·{\cos \mathrm{\θ}}_x)/\cos \left({\mathrm{\θ}}_x\right)\·L_y^{\′}/L_y+y_0^r\end{array}\right.---\left(3\right)$

Wherein: for the coordinate value of initial point in dot matrix under robot coordinate system;

L _x, L _yfor the physical length of dot matrix mid point line;

L' _x, L' _yfor the length of dot matrix mid point line under robot coordinate system;

θ _xfor the deflection angle of visual coordinate system axis of ordinates and robot coordinate system's axis of ordinates;

θ _yfor the deflection angle of visual coordinate system axis of ordinates and robot coordinate system's axis of ordinates;

4.2) to obtain on benchmark workpiece the coordinate a little under assembly robot coordinate system:

4.2.1) by benchmark workpiece movable under the station of assembly robot, adjustment robot end installing plate is parallel with benchmark workpiece, records robot end's anglec of rotation now;

4.2.2) in acquisition point matrix, the coordinate value of initial point under robot coordinate system ties up to the translational movement under robot coordinate system as visual coordinate;

4.2.3) angle between some line and the axis of ordinates of robot coordinate system arranged arbitrarily in the angle between some line and the axis of abscissas of robot coordinate system and dot matrix of going arbitrarily in dot matrix is got respectively; The algorithm computation vision coordinate system of averaging according to suing for peace and the deflection angle of robot coordinate system;

4.2.4) obtaining the conversion formula that visual coordinate is tied to robot coordinate system is:

$\left\{\begin{array}{c}{x}_1^r=x^r+L\·\cos ({\mathrm{\θ}}_1+{\mathrm{\θ}}_L)-L\·\cos ({\mathrm{\θ}}_1+{\mathrm{\θ}}_L+{\mathrm{\θ}}_0)\\ y_1^r=y^r+L\·\sin ({\mathrm{\θ}}_1+{\mathrm{\θ}}_L)-L\·\cos ({\mathrm{\θ}}_1+{\mathrm{\θ}}_L+{\mathrm{\θ}}_0)\end{array}\right.---\left(4\right)$

Wherein: L is the distance of alignment point in robot rotation axis points to robot end's installing plate;

θ _lfor the line of alignment point in robot end's rotation axis points and robot end's installing plate and the angle of visual coordinate system abscissa;

θ ₁for the deflection angle of visual coordinate system and robot coordinate system;

θ ₀for robot end's anglec of rotation;

5) robot picks up workpiece according to the coordinate of the workpiece obtained under robot coordinate system.

2. Robot Hand-eye localization method according to claim 1, is characterized in that: described image coordinate system is the upper left corner of the image taken is initial point, and the transverse direction of shooting image is axis of abscissas, and the longitudinal direction of shooting image is the coordinate system that axis of ordinates is set up.