CN101301711A - Automatic Welding and Positioning System for Reinforcing Plate of Container Top Corner - Google Patents

Abstract

The invention discloses an automatic welding and positioning system for the reinforcement plate of the top corner of a container, which is characterized in that it includes a camera, two lasers, and a computer; wherein the optical axis of the camera is installed vertically on the working platform, and the optical planes of the two lasers are parallel and bound together. Together, it is installed at an angle of 45° to the optical axis of the camera; the camera is connected to the computer through the image acquisition card, and the computer is connected to the welding robot through the data line, which can control the robot to perform corresponding operations. The automatic welding and positioning system for the reinforcing plate of the container top corner piece of the present invention can realize a fast and effective posture recognition and positioning method of the reinforcing plate at the initial point through a simple image processing algorithm and a simplified positioning processing method.

Description

Automatic Welding and Positioning System for Reinforcing Plate of Container Top Corner

technical field

The invention relates to the field of robot welding systems for welding container top corner fitting reinforcement plates, in particular to a positioning system used in the process of robot welding top corner fittings in this field.

Background technique

At present, 80% of the welding work in the container manufacturing plant adopts manual welding, including corrugated welding, direct welding, vertical welding, lap welding and so on. Due to the rough tooling, it is difficult to realize automatic welding. If the tooling is improved and fully automated, it will increase costs and reduce production speed. This is a way that the current highly competitive container production industry cannot and is unwilling to implement. However, driven by the increasing social interests of people-oriented and humanized management, workers are required to be freed from the harsh polluted environment while ensuring welding quality and welding efficiency. In order to solve this problem, Qingdao Branch of China Container Manufacturing Co., Ltd., the largest container manufacturer in the world, put forward a demand to study a special automatic welding robot for different welding types. Among them, the welding process of the reinforcement plate of the container top corner is simple, and the shape and size of the plate are known. However, due to the rough production environment and inaccurate positioning of the tooling, the position and posture of the reinforcement plate cannot be accurately positioned, and automatic welding cannot be realized. In addition, this One type of work requires reinforcement plates to be welded on the four corners of the container, which requires four workers to complete, resulting in a large labor force and high welding costs. Therefore, in order to meet the actual needs of production, it has become an urgent need to develop a set of positioning system for automatic welding of corner reinforcement plates with low cost, high positioning accuracy, easy to use, and can be popularized and applied.

Contents of the invention

The task of the present invention is to provide an automatic welding and positioning system for the reinforcement plate of the container top corner fitting. The system can realize a fast and effective positioning of the initial point of the reinforcement plate and gesture recognition through a simple image processing algorithm and a simplified positioning processing method. method.

Its technical solutions are:

An automatic welding and positioning system for the reinforcing plate of the container top corner piece, including a camera, two lasers, and a computer; the optical axis of the camera is installed vertically on the working platform, and the optical planes of the two lasers are parallel and bundled together, and are connected with the optical axis of the camera. The axis is installed at an angle of 45°; the camera is connected to the computer through the image acquisition card, and the computer is connected to the welding robot through the data line, which can control the robot to perform corresponding operations.

The above-mentioned cameras, double lasers, and blower are installed on the right-angle bracket; the blower is fixed on the right-angle bracket, and the nozzle faces the working platform; the camera is connected to the computer through the 1394 card, and the blower and the laser are directly connected to the power supply; the computer communicates with the welding machine through the RS232 protocol. Robot communication.

The image plane of the above-mentioned camera is parallel to the working plane, and the image plane and the working plane form a corresponding proportional relationship, and this proportion data is obtained through a standard rectangular template; when initially installing the equipment, first set the origin of the working coordinate system through the computer, and use the initial welding of the template The point space coordinate is the origin, the system takes the first welding target as the template, and the two outer sides of the template are the x-axis and y-axis of the coordinate system, and its image information is stored in the template library; during the online positioning process, the double laser stripe At the outer corner of the reinforcement board, the camera acquires the image of the reinforcement board with laser marking in real time, and obtains the image information of the initial point and its attitude information through computer processing; finally, the space of the real-time reinforcement board is obtained through the conversion relationship between the image information and the space information Initial point position and pose.

The above-mentioned automatic welding positioning system for the reinforcement plate of the container top corner piece includes the establishment method of the corresponding relationship between the spatial object and its image, the establishment method of the corresponding relationship between the online object information and the reference template, the acquisition method of the object depth information and the online object information. method of obtaining;

The way to establish the corresponding relationship between the above space objects and their images is as follows:

The corresponding proportional relationship between the image plane and the working plane is:

$\left[\begin{array}{c}\mathrm{<span\; class="notranslate">\; x</span>}\\ \mathrm{<span\; class="notranslate">\; the\; y</span>}\end{array}\right]<span\; class="notranslate">\; =</span>\left[\begin{array}{cc}\mathrm{<span\; class="notranslate">\; a</span>}& \mathrm{<span\; class="notranslate">\; b</span>}\end{array}\right]\left[\begin{array}{c}\mathrm{<span\; class="notranslate">\; u</span>}\\ \mathrm{<span\; class="notranslate">\; v</span>}\end{array}\right]<span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; )</span>$

Use a camera to capture a rectangular template image of known size, obtain the length and width of the rectangular template in pixels from the image, and then obtain the ratio of the image space to the physical space by a=x/u, b=y/v coefficient a, b;

The way to establish the corresponding relationship between the above-mentioned online target information and the reference template is as follows:

与x轴方向的夹角记为θ，

向量记为t，空间点 $\{({\tilde{x}}_1,{\tilde{y}}_1),({\tilde{x}}_2,{\tilde{y}}_2),...,({\tilde{x}}_6,{\tilde{y}}_6)\}$ 在坐标系x-O-y下的坐标为{(x₁，y₁)，(x₂，y₂)，...，(x₆，y₆)}，可通过下式获得：In the working plane coordinate system, the 6 points of the reinforced plate image obtained online are transformed into space coordinate points by formula (1) $\{({\tilde{x}}_1,{\widetilde{\mathrm{the\; y}}}_1),({\tilde{x}}_2,{\widetilde{\mathrm{the\; y}}}_2),...,({\tilde{x}}_6,{\widetilde{\mathrm{the\; y}}}_6)\},$ Will

The included angle with the x-axis direction is denoted as θ,

The vector is denoted as t, the space point $\{({\tilde{x}}_1,{\widetilde{\mathrm{the\; y}}}_1),({\tilde{x}}_2,{\widetilde{\mathrm{the\; y}}}_2),...,({\tilde{x}}_6,{\widetilde{\mathrm{the\; y}}}_6)\}$ The coordinates in the coordinate system xOy are {(x ₁ , y ₁ ), (x ₂ , y ₂ ), ..., (x ₆ , y ₆ )}, which can be obtained by the following formula:

$\left[\begin{array}{c}\mathrm{<span\; class="notranslate">\; x</span>}\\ \mathrm{<span\; class="notranslate">\; the\; y</span>}\end{array}\right]<span\; class="notranslate">\; =</span>\left[\begin{array}{cc}\mathrm{<span\; class="notranslate">\; cos</span>}\mathrm{<span\; class="notranslate">\; \&theta;</span>}& \mathrm{<span\; class="notranslate">\; sin</span>}\mathrm{<span\; class="notranslate">\; \&theta;</span>}\\ <span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; sin</span>}\mathrm{<span\; class="notranslate">\; \&theta;</span>}& \mathrm{<span\; class="notranslate">\; cos</span>}\mathrm{<span\; class="notranslate">\; \&theta;</span>}\end{array}\right]\left[\begin{array}{c}\stackrel{<span\; class="notranslate">\; ~</span>}{\mathrm{<span\; class="notranslate">\; x</span>}}\\ \stackrel{<span\; class="notranslate">\; ~</span>}{\mathrm{<span\; class="notranslate">\; the\; y</span>}}\end{array}\right]<span\; class="notranslate">\; +</span>\mathrm{<span\; class="notranslate">\; t</span>}<span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 2</span>}<span\; class="notranslate">\; )</span>$

After calculating (x ₁ , y ₁ ) and (x ₂ , y ₂ ), determine the positions of other points according to the direction and known dimensions of the reinforcement plate;

The method of obtaining the above-mentioned target depth information is as follows:

Assuming three parallel planes, plane A is the image plane, planes B and C are space planes of different heights, the size of the image formed by the same object at different heights is different, the specific derivation process is as follows:

$\frac{<span\; class="notranslate">\; |</span><span\; class="notranslate">\; |</span>\stackrel{<span\; class="notranslate">\; \&Right\; Arrow;</span>}{{\mathrm{<span\; class="notranslate">\; OO</span>}}_{\mathrm{<span\; class="notranslate">\; 1</span>}}}<span\; class="notranslate">\; |</span><span\; class="notranslate">\; |</span>}{<span\; class="notranslate">\; |</span><span\; class="notranslate">\; |</span>\stackrel{<span\; class="notranslate">\; \&Right\; Arrow;</span>}{{\mathrm{<span\; class="notranslate">\; OO</span>}}_{\mathrm{<span\; class="notranslate">\; 3</span>}}}<span\; class="notranslate">\; |</span><span\; class="notranslate">\; |</span>}<span\; class="notranslate">\; =</span>\frac{<span\; class="notranslate">\; |</span><span\; class="notranslate">\; |</span>\stackrel{<span\; class="notranslate">\; \&Right\; Arrow;</span>}{{\mathrm{<span\; class="notranslate">\; a</span>}}_{\mathrm{<span\; class="notranslate">\; 1</span>}}{\mathrm{<span\; class="notranslate">\; b</span>}}_{\mathrm{<span\; class="notranslate">\; 1</span>}}}<span\; class="notranslate">\; |</span><span\; class="notranslate">\; |</span>}{<span\; class="notranslate">\; |</span><span\; class="notranslate">\; |</span>\stackrel{<span\; class="notranslate">\; \&Right\; Arrow;</span>}{{\mathrm{<span\; class="notranslate">\; c</span>}}_{\mathrm{<span\; class="notranslate">\; 1</span>}}{\mathrm{<span\; class="notranslate">\; d</span>}}_{\mathrm{<span\; class="notranslate">\; 1</span>}}}<span\; class="notranslate">\; |</span><span\; class="notranslate">\; |</span>}<span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 3</span>}<span\; class="notranslate">\; )</span>$

$\frac{<span\; class="notranslate">\; |</span><span\; class="notranslate">\; |</span>\stackrel{<span\; class="notranslate">\; \&Right\; Arrow;</span>}{{\mathrm{<span\; class="notranslate">\; OO</span>}}_{\mathrm{<span\; class="notranslate">\; 1</span>}}}<span\; class="notranslate">\; |</span><span\; class="notranslate">\; |</span>}{<span\; class="notranslate">\; |</span><span\; class="notranslate">\; |</span>\stackrel{<span\; class="notranslate">\; \&Right\; Arrow;</span>}{{\mathrm{<span\; class="notranslate">\; <figure-callout\; id="2"\; label="OO"\; filenames="A20081001697300091.png"\; state="\{\{state\}\}">OO</figure-callout></span>}}_{\mathrm{<span\; class="notranslate">\; 2</span>}}}<span\; class="notranslate">\; |</span><span\; class="notranslate">\; |</span>}<span\; class="notranslate">\; =</span>\frac{<span\; class="notranslate">\; |</span><span\; class="notranslate">\; |</span>\stackrel{<span\; class="notranslate">\; \&Right\; Arrow;</span>}{{\mathrm{<span\; class="notranslate">\; a</span>}}_{\mathrm{<span\; class="notranslate">\; 2</span>}}{\mathrm{<span\; class="notranslate">\; <figure-callout\; id="2"\; label="b"\; filenames="A20081001697300091.png"\; state="\{\{state\}\}">b</figure-callout></span>}}_{\mathrm{<span\; class="notranslate">\; 2</span>}}}<span\; class="notranslate">\; |</span><span\; class="notranslate">\; |</span>}{<span\; class="notranslate">\; |</span><span\; class="notranslate">\; |</span>\stackrel{<span\; class="notranslate">\; \&Right\; Arrow;</span>}{{\mathrm{<span\; class="notranslate">\; c</span>}}_{\mathrm{<span\; class="notranslate">\; 2</span>}}{\mathrm{<span\; class="notranslate">\; <figure-callout\; id="2"\; label="d"\; filenames="A20081001697300091.png"\; state="\{\{state\}\}">d</figure-callout></span>}}_{\mathrm{<span\; class="notranslate">\; 2</span>}}}<span\; class="notranslate">\; |</span><span\; class="notranslate">\; |</span>}<span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 4</span>}<span\; class="notranslate">\; )</span>$

的是事先已知的深度信息，

和

是同一目标的长度，由公式(3)和(4)可得：in

is the depth information known in advance,

and

is the length of the same target, which can be obtained from formulas (3) and (4):

$\frac{<span\; class="notranslate">\; |</span><span\; class="notranslate">\; |</span>\stackrel{<span\; class="notranslate">\; \&Right\; Arrow;</span>}{{\mathrm{<span\; class="notranslate">\; OO</span>}}_{\mathrm{<span\; class="notranslate">\; 3</span>}}}<span\; class="notranslate">\; |</span><span\; class="notranslate">\; |</span>}{<span\; class="notranslate">\; |</span><span\; class="notranslate">\; |</span>\stackrel{<span\; class="notranslate">\; \&Right\; Arrow;</span>}{{\mathrm{<span\; class="notranslate">\; <figure-callout\; id="2"\; label="OO"\; filenames="A20081001697300091.png"\; state="\{\{state\}\}">OO</figure-callout></span>}}_{\mathrm{<span\; class="notranslate">\; 2</span>}}}<span\; class="notranslate">\; |</span><span\; class="notranslate">\; |</span>}<span\; class="notranslate">\; =</span>\frac{<span\; class="notranslate">\; |</span><span\; class="notranslate">\; |</span>\stackrel{<span\; class="notranslate">\; \&Right\; Arrow;</span>}{{\mathrm{<span\; class="notranslate">\; a</span>}}_{\mathrm{<span\; class="notranslate">\; 2</span>}}{\mathrm{<span\; class="notranslate">\; <figure-callout\; id="2"\; label="b"\; filenames="A20081001697300091.png"\; state="\{\{state\}\}">b</figure-callout></span>}}_{\mathrm{<span\; class="notranslate">\; 2</span>}}}<span\; class="notranslate">\; |</span><span\; class="notranslate">\; |</span>}{<span\; class="notranslate">\; |</span><span\; class="notranslate">\; |</span>\stackrel{<span\; class="notranslate">\; \&Right\; Arrow;</span>}{{\mathrm{<span\; class="notranslate">\; a</span>}}_{\mathrm{<span\; class="notranslate">\; 1</span>}}{\mathrm{<span\; class="notranslate">\; b</span>}}_{\mathrm{<span\; class="notranslate">\; 1</span>}}}<span\; class="notranslate">\; |</span><span\; class="notranslate">\; |</span>}<span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 5</span>}<span\; class="notranslate">\; )</span>$

For the convenience of calculation, formula (5) is transformed into:

$<span\; class="notranslate">\; |</span><span\; class="notranslate">\; |</span>\stackrel{<span\; class="notranslate">\; \&Right\; Arrow;</span>}{{\mathrm{<span\; class="notranslate">\; OO</span>}}_{\mathrm{<span\; class="notranslate">\; 3</span>}}}<span\; class="notranslate">\; |</span><span\; class="notranslate">\; |</span><span\; class="notranslate">\; =</span>\frac{\sqrt{{\mathrm{<span\; class="notranslate">\; S</span>}}_{\mathrm{<span\; class="notranslate">\; 2</span>}}}}{\sqrt{{\mathrm{<span\; class="notranslate">\; S</span>}}_{\mathrm{<span\; class="notranslate">\; 1</span>}}}}<span\; class="notranslate">\; |</span><span\; class="notranslate">\; |</span>\stackrel{<span\; class="notranslate">\; \&Right\; Arrow;</span>}{{\mathrm{<span\; class="notranslate">\; <figure-callout\; id="2"\; label="OO"\; filenames="A20081001697300091.png"\; state="\{\{state\}\}">OO</figure-callout></span>}}_{\mathrm{<span\; class="notranslate">\; 2</span>}}}<span\; class="notranslate">\; |</span><span\; class="notranslate">\; |</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 6</span>}<span\; class="notranslate">\; )</span>$

为参考模板的深度信息，

为在线检测到的目标深度信息。 _S1 is the area of the small rectangle in the image plane and _S2 is the area of the large rectangle. Using this formula to extend the target shape to any polygon, this relationship exists in any posture, and the calculation is very convenient. so

For the depth information of the reference template,

is the target depth information detected online.

The above online target information can be obtained as follows:

The above laser marks are two red laser stripes, using the turning points formed by the two laser surfaces on the reinforcing plate to obtain the position of the initial welding point; the specific process is to first extract the R image of the RGB color space, and then select the threshold Value and denoising processing. After thinning processing, the single-pixel image of the light stripe is obtained. Finally, the boundary points formed by the stripes on the reinforcing plate are obtained by using the column scanning method. The four boundary points are used to fit a straight line and find their intersection point, that is, Image coordinates (u ₁ , v ₁ ) of the welding initial point, and then obtain the first point (u ₂ , v ₂ ) sorted clockwise starting from (u ₁ , v ₁ ) through Hough transform.

The automatic welding and positioning system for the reinforcing plate of the container top corner piece of the present invention can realize a fast and effective posture recognition and positioning method of the reinforcing plate at the initial point through a simple image processing algorithm and a simplified positioning processing method. The system can determine the size and thickness of the reinforcement plate through the box type parameters set by the system to complete the task with an effective algorithm. In addition, since the size and thickness of the reinforcing plate are known, the automatic welding task can be realized once the initial point position and the posture of the reinforcing plate are determined.

Description of drawings

Fig. 1 is a schematic schematic diagram of an embodiment of the present invention, and also shows other items related to the present invention.

FIG. 2 shows a corresponding relationship between spatial objects and images related to the present invention.

FIG. 3 shows a corresponding relationship between online object information and reference templates related to the present invention.

Fig. 4 shows an acquisition method of target depth information related to the present invention.

Figures 5a-5e show a method of acquiring online target information related to the present invention

The present invention is described below in conjunction with accompanying drawing:

Detailed ways

As shown in FIG. 1 , an automatic welding positioning system for a reinforcing plate of a container top corner fitting includes a camera 1 , two lasers 2 , a blower 7 , a computer 3 or other suitable computing devices. The specific connection method is that the camera, double lasers, and blower are installed on the right-angle bracket 8, wherein the optical axis of the camera is required to be installed vertically on the working platform, the optical planes of the double lasers are parallel and bound together, and installed at an angle of 45° with the optical axis of the camera; The tube 7 is fixed on the right-angle bracket 8, and the nozzle faces the working platform; the camera is connected to the computer through the 1394 card 11; the blower is directly connected to the laser and the power supply; the computer 3 communicates with the welding robot 4 through the RS232 protocol, and can control the robot to perform corresponding operations (Task).

The image plane of the above-mentioned camera is parallel to the working plane, and the image plane and the working plane form a corresponding proportional relationship, and this proportion data is obtained through a standard rectangular template; when initially installing the equipment, first set the origin of the working coordinate system through the computer, and use the initial welding of the template The point space coordinate is the origin, the system takes the first welding target as the template, and the two outer sides of the template are the x-axis and y-axis of the coordinate system, and its image information is stored in the template library; during the online positioning process, the double laser stripe At the outer corner 10 of the reinforcing plate 5 on the top plate 6 of the container, the camera acquires the image of the reinforcing plate 5 with the laser mark in real time, and obtains the image information of the initial point 9 and its posture information through computer processing; finally, through the image information and space The transformation relation of information obtains the space initial point position and attitude of the real-time reinforced plate.

垂直空间平面x-O″-y和图像平面x-O′-y，并存在如下关系：Referring to Figure 2, since the optical axis is perpendicular to the working platform, that is

The vertical space plane xO″-y and the image plane xO′-y have the following relationship:

$\left[\begin{array}{c}\mathrm{<span\; class="notranslate">\; x</span>}\\ \mathrm{<span\; class="notranslate">\; the\; y</span>}\end{array}\right]<span\; class="notranslate">\; =</span>\left[\begin{array}{cc}\mathrm{<span\; class="notranslate">\; a</span>}& \mathrm{<span\; class="notranslate">\; b</span>}\end{array}\right]\left[\begin{array}{c}\mathrm{<span\; class="notranslate">\; u</span>}\\ \mathrm{<span\; class="notranslate">\; v</span>}\end{array}\right]<span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; )</span>$

Use a camera to capture a rectangular template image of known size, obtain the length and width of the rectangular template in pixels from the image, and then obtain the ratio of the image space to the physical space through a=x/u, b=y/v Coefficients a, b.

向量记为t，空间点 $\{({\tilde{x}}_1,{\tilde{y}}_1),({\tilde{x}}_2,{\tilde{y}}_2),...,({\tilde{x}}_6,{\tilde{y}}_6)\}$ 在坐标系x-O-y下的坐标为{(x₁，y₁)，(x₂，y₂)，...，(x₆，y₆)}，可通过下式获得：Referring to Figure 3, in the coordinate system of the working plane, 12 is the reference template, 13 is the online reinforcement plate, O is the origin of space coordinates (0, 0, 0), and the two outer boundaries are the coordinates x-axis and y-axis respectively, and the reference template The image information of the template is stored in the template library in advance. Transform the 6 points of the reinforced plate image obtained online into spatial coordinate points by formula (1) $\{({\tilde{x}}_1,{\widetilde{\mathrm{the\; y}}}_1),({\tilde{x}}_2,{\widetilde{\mathrm{the\; y}}}_2),...,({\tilde{x}}_6,{\widetilde{\mathrm{the\; y}}}_6)\}.$ Will The included angle with the x-axis direction is denoted as θ,

The vector is denoted as t, the space point $\{({\tilde{x}}_1,{\widetilde{\mathrm{the\; y}}}_1),({\tilde{x}}_2,{\widetilde{\mathrm{the\; y}}}_2),...,({\tilde{x}}_6,{\widetilde{\mathrm{the\; y}}}_6)\}$ The coordinates in the coordinate system xOy are {(x ₁ , y ₁ ), (x ₂ , y ₂ ), ..., (x ₆ , y ₆ )}, which can be obtained by the following formula:

$\left[\begin{array}{c}\mathrm{<span\; class="notranslate">\; x</span>}\\ \mathrm{<span\; class="notranslate">\; the\; y</span>}\end{array}\right]<span\; class="notranslate">\; =</span>\left[\begin{array}{cc}\mathrm{<span\; class="notranslate">\; cos</span>}\mathrm{<span\; class="notranslate">\; \&theta;</span>}& \mathrm{<span\; class="notranslate">\; sin</span>}\mathrm{<span\; class="notranslate">\; \&theta;</span>}\\ <span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; sin</span>}\mathrm{<span\; class="notranslate">\; \&theta;</span>}& \mathrm{<span\; class="notranslate">\; cos</span>}\mathrm{<span\; class="notranslate">\; \&theta;</span>}\end{array}\right]\left[\begin{array}{c}\stackrel{<span\; class="notranslate">\; ~</span>}{\mathrm{<span\; class="notranslate">\; x</span>}}\\ \stackrel{<span\; class="notranslate">\; ~</span>}{\mathrm{<span\; class="notranslate">\; the\; y</span>}}\end{array}\right]<span\; class="notranslate">\; +</span>\mathrm{<span\; class="notranslate">\; t</span>}<span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 2</span>}<span\; class="notranslate">\; )</span>$

In practical applications, we only need to find out (x ₁ , y ₁ ) and (x ₂ , y ₂ ), that is, the position of the initial welding point and the direction of the reinforcement plate, and determine and arrange the other 4 points according to the order of known dimensions.

Referring to Figure 4, there are three parallel planes, plane A is the image plane, planes B and C are space planes of different heights, the size of the image formed by the same object at different heights is different, the figure shows that the depth information of the target is related to the formed The area of the image is related, and the specific derivation process is as follows:

$\frac{<span\; class="notranslate">\; |</span><span\; class="notranslate">\; |</span>\stackrel{<span\; class="notranslate">\; \&Right\; Arrow;</span>}{{\mathrm{<span\; class="notranslate">\; OO</span>}}_{\mathrm{<span\; class="notranslate">\; 1</span>}}}<span\; class="notranslate">\; |</span><span\; class="notranslate">\; |</span>}{<span\; class="notranslate">\; |</span><span\; class="notranslate">\; |</span>\stackrel{<span\; class="notranslate">\; \&Right\; Arrow;</span>}{{\mathrm{<span\; class="notranslate">\; OO</span>}}_{\mathrm{<span\; class="notranslate">\; 3</span>}}}<span\; class="notranslate">\; |</span><span\; class="notranslate">\; |</span>}<span\; class="notranslate">\; =</span>\frac{<span\; class="notranslate">\; |</span><span\; class="notranslate">\; |</span>\stackrel{<span\; class="notranslate">\; \&Right\; Arrow;</span>}{{\mathrm{<span\; class="notranslate">\; a</span>}}_{\mathrm{<span\; class="notranslate">\; 1</span>}}{\mathrm{<span\; class="notranslate">\; b</span>}}_{\mathrm{<span\; class="notranslate">\; 1</span>}}}<span\; class="notranslate">\; |</span><span\; class="notranslate">\; |</span>}{<span\; class="notranslate">\; |</span><span\; class="notranslate">\; |</span>\stackrel{<span\; class="notranslate">\; \&Right\; Arrow;</span>}{{\mathrm{<span\; class="notranslate">\; c</span>}}_{\mathrm{<span\; class="notranslate">\; 1</span>}}{\mathrm{<span\; class="notranslate">\; d</span>}}_{\mathrm{<span\; class="notranslate">\; 1</span>}}}<span\; class="notranslate">\; |</span><span\; class="notranslate">\; |</span>}<span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 3</span>}<span\; class="notranslate">\; )</span>$

$\frac{<span\; class="notranslate">\; |</span><span\; class="notranslate">\; |</span>\stackrel{<span\; class="notranslate">\; \&Right\; Arrow;</span>}{{\mathrm{<span\; class="notranslate">\; OO</span>}}_{\mathrm{<span\; class="notranslate">\; 1</span>}}}<span\; class="notranslate">\; |</span><span\; class="notranslate">\; |</span>}{<span\; class="notranslate">\; |</span><span\; class="notranslate">\; |</span>\stackrel{<span\; class="notranslate">\; \&Right\; Arrow;</span>}{{\mathrm{<span\; class="notranslate">\; <figure-callout\; id="2"\; label="OO"\; filenames="A20081001697300091.png"\; state="\{\{state\}\}">OO</figure-callout></span>}}_{\mathrm{<span\; class="notranslate">\; 2</span>}}}<span\; class="notranslate">\; |</span><span\; class="notranslate">\; |</span>}<span\; class="notranslate">\; =</span>\frac{<span\; class="notranslate">\; |</span><span\; class="notranslate">\; |</span>\stackrel{<span\; class="notranslate">\; \&Right\; Arrow;</span>}{{\mathrm{<span\; class="notranslate">\; a</span>}}_{\mathrm{<span\; class="notranslate">\; 2</span>}}{\mathrm{<span\; class="notranslate">\; <figure-callout\; id="2"\; label="b"\; filenames="A20081001697300091.png"\; state="\{\{state\}\}">b</figure-callout></span>}}_{\mathrm{<span\; class="notranslate">\; 2</span>}}}<span\; class="notranslate">\; |</span><span\; class="notranslate">\; |</span>}{<span\; class="notranslate">\; |</span><span\; class="notranslate">\; |</span>\stackrel{<span\; class="notranslate">\; \&Right\; Arrow;</span>}{{\mathrm{<span\; class="notranslate">\; c</span>}}_{\mathrm{<span\; class="notranslate">\; 2</span>}}{\mathrm{<span\; class="notranslate">\; <figure-callout\; id="2"\; label="d"\; filenames="A20081001697300091.png"\; state="\{\{state\}\}">d</figure-callout></span>}}_{\mathrm{<span\; class="notranslate">\; 2</span>}}}<span\; class="notranslate">\; |</span><span\; class="notranslate">\; |</span>}<span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 4</span>}<span\; class="notranslate">\; )</span>$

是同一目标的长度，是相等的。由公式(3)和(4)可得：in is the depth information known in advance, and

are the lengths of the same target and are equal. From formulas (3) and (4), we can get:

$\frac{<span\; class="notranslate">\; |</span><span\; class="notranslate">\; |</span>\stackrel{<span\; class="notranslate">\; \&Right\; Arrow;</span>}{{\mathrm{<span\; class="notranslate">\; OO</span>}}_{\mathrm{<span\; class="notranslate">\; 3</span>}}}<span\; class="notranslate">\; |</span><span\; class="notranslate">\; |</span>}{<span\; class="notranslate">\; |</span><span\; class="notranslate">\; |</span>\stackrel{<span\; class="notranslate">\; \&Right\; Arrow;</span>}{{\mathrm{<span\; class="notranslate">\; <figure-callout\; id="2"\; label="OO"\; filenames="A20081001697300091.png"\; state="\{\{state\}\}">OO</figure-callout></span>}}_{\mathrm{<span\; class="notranslate">\; 2</span>}}}<span\; class="notranslate">\; |</span><span\; class="notranslate">\; |</span>}<span\; class="notranslate">\; =</span>\frac{<span\; class="notranslate">\; |</span><span\; class="notranslate">\; |</span>\stackrel{<span\; class="notranslate">\; \&Right\; Arrow;</span>}{{\mathrm{<span\; class="notranslate">\; a</span>}}_{\mathrm{<span\; class="notranslate">\; 2</span>}}{\mathrm{<span\; class="notranslate">\; <figure-callout\; id="2"\; label="b"\; filenames="A20081001697300091.png"\; state="\{\{state\}\}">b</figure-callout></span>}}_{\mathrm{<span\; class="notranslate">\; 2</span>}}}<span\; class="notranslate">\; |</span><span\; class="notranslate">\; |</span>}{<span\; class="notranslate">\; |</span><span\; class="notranslate">\; |</span>\stackrel{<span\; class="notranslate">\; \&Right\; Arrow;</span>}{{\mathrm{<span\; class="notranslate">\; a</span>}}_{\mathrm{<span\; class="notranslate">\; 1</span>}}{\mathrm{<span\; class="notranslate">\; b</span>}}_{\mathrm{<span\; class="notranslate">\; 1</span>}}}<span\; class="notranslate">\; |</span><span\; class="notranslate">\; |</span>}<span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 5</span>}<span\; class="notranslate">\; )</span>$

For the convenience of calculation, formula (5) is transformed into:

$<span\; class="notranslate">\; |</span><span\; class="notranslate">\; |</span>\stackrel{<span\; class="notranslate">\; \&Right\; Arrow;</span>}{{\mathrm{<span\; class="notranslate">\; OO</span>}}_{\mathrm{<span\; class="notranslate">\; 3</span>}}}<span\; class="notranslate">\; |</span><span\; class="notranslate">\; |</span><span\; class="notranslate">\; =</span>\frac{\sqrt{{\mathrm{<span\; class="notranslate">\; S</span>}}_{\mathrm{<span\; class="notranslate">\; 2</span>}}}}{\sqrt{{\mathrm{<span\; class="notranslate">\; S</span>}}_{\mathrm{<span\; class="notranslate">\; 1</span>}}}}<span\; class="notranslate">\; |</span><span\; class="notranslate">\; |</span>\stackrel{<span\; class="notranslate">\; \&Right\; Arrow;</span>}{{\mathrm{<span\; class="notranslate">\; <figure-callout\; id="2"\; label="OO"\; filenames="A20081001697300091.png"\; state="\{\{state\}\}">OO</figure-callout></span>}}_{\mathrm{<span\; class="notranslate">\; 2</span>}}}<span\; class="notranslate">\; |</span><span\; class="notranslate">\; |</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 6</span>}<span\; class="notranslate">\; )</span>$

为参考模板的深度信息，为在线检测到的目标深度信息。 _S1 is the area of the small rectangle in the image plane and _S2 is the area of the large rectangle. Using this formula to extend the target shape to any polygon, this relationship exists in any posture, and the calculation is very convenient. so

For the depth information of the reference template, is the target depth information detected online.

Referring to Figures 5a-5e, in order to improve image processing efficiency, two red laser stripes are marked on the reinforcement plate, and the initial welding point is obtained by using the turning points formed by the two laser surfaces on the reinforcement plate, as shown in Figure 5a. The specific method is to first extract the R image in the RGB color space, as shown in Figure 5b; then select a reasonable threshold to perform binarization and denoising processing on it, as shown in Figure 5c, and obtain the single-pixel image of the light bar after thinning processing, as shown in Fig. 5d, 5e; Finally, use the column scanning method to obtain the boundary points formed by the stripes on the reinforcement plate, fit a straight line through the four boundary points and find their intersection point, that is, the image coordinates (u ₁ , v ₁ ) of the welding initial point, and then Obtain the first point (u ₂ , v ₂ ) sorted clockwise starting from (u ₁ , v ₁ ) through Hough transform, and compare the direction angle between (u ₁ , v ₁ ) and (u ₂ , v ₂ ) Denote the pose of the stiffener.

The column scan in this method obtains the boundary points, and the premise that the clockwise sorting can be realized is that in this application, the range of the pose of the stiffener is small (displacement<5mm, angle<5°), therefore, the camera can be installed in a place suitable for image calculation The position of , that is, the image u-axis is approximately parallel to the laser stripes.

The present invention collects images before welding, so it will not be affected by the arc welding light. When the welding robot changes direction or stops, it first ventilates through the blower pipe to blow away the smoke and dust in front of the camera lens, and then collects images .

Claims (4)

1, a kind of positioning system for automatic soldering of container apex angle piece reinforcing plate, feature are to comprise a video camera, two laser instruments, and a computer; Wherein the vertical workbench of camera optical axis is installed, and the optical plane of twin-laser is parallel and bundle, and installs with camera optical axis angle at 45; Video camera is connected with computer by image pick-up card, and the computer expert crosses data circuit and is connected with welding robot, can control robot and carry out corresponding operating.

2, positioning system for automatic soldering of container apex angle piece reinforcing plate according to claim 1 is characterized in that: described video camera, twin-laser, blower are installed on the right angle rack; Blower is fixed on the right angle rack, and the mouth of pipe is towards workbench; Video camera is connected with computer by 1394 cards, and blower and laser instrument directly connect power supply; The computer expert crosses the RS232 agreement and communicates by letter with welding robot.

3, positioning system for automatic soldering of container apex angle piece reinforcing plate according to claim 1, it is characterized in that: the plane of delineation of described video camera is parallel with working face, the plane of delineation forms corresponding proportionate relationship with working face, and obtains this ratio data by the standard rectangular template; During initial erection unit, at first by computer installation work coordinate system initial point, be initial point with the welding initial point space coordinates of template, native system is a template with first welding target, two outsides of template are respectively coordinate system x axle and y axle, and deposit its image information in ATL; In online position fixing process, two laser stripes are beaten at place, stiffener exterior angle, and video camera obtains the stiffener image that has laser labelling in real time, and obtain image information and its attitude information of initial point by Computer Processing; Obtain the initial point position, space and the attitude of real-time stiffener at last by the transformational relation of image information and spatial information.

4, positioning system for automatic soldering of container apex angle piece reinforcing plate according to claim 3, it is characterized in that: described positioning system for automatic soldering of container apex angle piece reinforcing plate comprises between extraterrestrial target and its image corresponding relation corresponding relation obtain manner of cube formula, target depth information and the obtain manner of online target information really between cube formula, online target information and the reference template really;

Cube formula is as follows really for corresponding relation between above-mentioned extraterrestrial target and its image:

The plane of delineation forms corresponding proportionate relationship formula with working face:

$\left[\begin{array}{c}x\\ y\end{array}\right]=\left[\begin{array}{cc}a& b\end{array}\right]\left[\begin{array}{c}u\\ v\end{array}\right]---\left(1\right)$

Utilize the rectangle template image of a known dimensions of camera acquisition, obtaining this rectangle template from image is the length and width size of unit with the pixel, then by a=x/u, and b=y/v, the proportionality coefficient a of acquisition image space and physical space, b;

Cube formula is as follows really for corresponding relation between above-mentioned online target information and the reference template:

In the working face coordinate system, 6 points of the stiffener image of online acquisition are converted into the space coordinates point by formula (1) Will

Be designated as θ with the axial angle of x,

Vector is designated as t, spatial point

Coordinate under coordinate system x-O-y is { (x ₁, y ₁), (x ₂, y ₂) ..., (x ₆, y ₆), can obtain by following formula:

$\left[\begin{array}{c}x\\ y\end{array}\right]=\left[\begin{array}{cc}\cos \mathrm{\&theta;}& \sin \mathrm{\&theta;}\\ -\sin \mathrm{\&theta;}& \cos \mathrm{\&theta;}\end{array}\right]\left[\begin{array}{c}\tilde{x}\\ \tilde{y}\end{array}\right]+t---\left(2\right)$

Obtain (x ₁, y ₁) and (x ₂, y ₂) after, determine other each point position according to the direction and the known dimensions of stiffener;

The obtain manner of above-mentioned target depth information is as follows:

If three parallel planes, plane A are the plane of delineation, plane B and C are the space planes of differing heights, and the formed image of same object of differing heights varies in size, and concrete derivation is as follows:

$\frac{\left|\right|\stackrel{\&RightArrow;}{{\mathrm{OO}}_1}\left|\right|}{\left|\right|\stackrel{\&RightArrow;}{{\mathrm{OO}}_3}\left|\right|}=\frac{\left|\right|\stackrel{\&RightArrow;}{a_1{b}_1}\left|\right|}{\left|\right|\stackrel{\&RightArrow;}{c_1{d}_1}\left|\right|}---\left(3\right)$

$\frac{\left|\right|\stackrel{\&RightArrow;}{{\mathrm{OO}}_1}\left|\right|}{\left|\right|\stackrel{\&RightArrow;}{{\mathrm{OO}}_2}\left|\right|}=\frac{\left|\right|\stackrel{\&RightArrow;}{a_2{b}_2}\left|\right|}{\left|\right|\stackrel{\&RightArrow;}{c_2{d}_2}\left|\right|}---\left(4\right)$

Wherein

Be prior known depth information,

With

Be the length of same target, can get by formula (3) and (4):

$\frac{\left|\right|\stackrel{\&RightArrow;}{{\mathrm{OO}}_3}\left|\right|}{\left|\right|\stackrel{\&RightArrow;}{{\mathrm{OO}}_2}\left|\right|}=\frac{\left|\right|\stackrel{\&RightArrow;}{a_2{b}_2}\left|\right|}{\left|\right|\stackrel{\&RightArrow;}{a_1{b}_1}\left|\right|}---\left(5\right)$

For convenience of calculation, formula (5) is converted into:

$\left|\right|\stackrel{\&RightArrow;}{{\mathrm{OO}}_3}\left|\right|=\frac{\sqrt{S_2}}{\sqrt{S_1}}\left|\right|\stackrel{\&RightArrow;}{{\mathrm{OO}}_2}\left|\right|---\left(6\right)$

S ₁Be the area of the medium and small rectangle of the plane of delineation, S ₂It is the area of large rectangle.Utilize this formula that target shape is generalized to arbitrary polygon, all there is this relation in attitude arbitrarily, calculates very convenient.Like this

Be the depth information of reference template,

Be online detected target depth information.

The obtain manner of above-mentioned online target information is as follows:

Above-mentioned laser labelling is two red laser stripeds, and the turning point that utilizes two lasing areas to form on stiffener obtains welding initial point position; Detailed process is, at first extract the R image of RGB color space, selected threshold is carried out binaryzation, denoising to it then, after thinning processing, obtain single pixel map of striation, utilize the column scan method to obtain the boundary point that striped forms at last on stiffener, fit straight line and ask its intersection point by four boundary points, promptly weld the image coordinate (u of initial point ₁, v ₁), obtain with (u by Hough transformation again ₁, v ₁) be starting point first point (u of ordering clockwise ₂, v ₂).

Images