CN2740335Y - Weld tracking visual sensor based on laser structural light - Google Patents

Abstract

The utility model relates to the technical field of robots, which is a welding seam tracking visual sensor based on laser structured light, adopts laser as a light source of structured light, and is composed of a laser generator, a camera, a monitor and a processing unit. Use the laser beam to form a laser plane through a cylindrical mirror, and project it onto the workpiece to generate laser stripes. The visual sensor collects the laser image of the weld seam, calculates the weld seam position through image processing, and converts it into an analog signal or wireless communication output to achieve control welding. The robot realizes the purpose of automatic seam tracking. Different welding seam image processing modules have been developed for various types of welds, and different processes such as one-time welds and multiple welds, so as to meet the welding seam tracking requirements of various types and processes. The utility model can improve the measurement accuracy of the welding seam position, increase the anti-interference ability of the welding seam tracking, and has the versatility and flexibility to adapt to various welding seam tracking systems.

Description

A welding seam tracking vision sensor based on laser structured light

Technical field:

The invention belongs to the visual measurement sensing device in the field of robots, in particular to a visual sensing device which can recognize the position of the welding seam in the welding process through visual images and guide the welding torch to correct the deviation.

technical background:

At present, the control methods used in welding industrial production generally use manual control or robot teaching or offline programming for path planning and motion programming. During the welding process, the preset actions are simply repeated. In existing seam tracking solutions, mechanical, electromagnetic, and visual sensors are generally used to provide seam information, which is processed by the system to realize welding torch position control. However, such weld seam tracking systems are generally closed, and are sensors and processing algorithms developed for specific systems, lacking general, flexible and independent welding sensors to provide weld seam position information for general welding robots or controllers.

Invention content:

In order to overcome the shortcomings of the existing seam tracking system, the utility model provides a seam tracking visual sensor. The sensor can adapt to various types of welds and different processes of one-time welds and multiple welds; collect laser weld images, extract the position of weld feature points; and provide analog output interfaces and network interfaces, and the sensor can be set up on the network to work Parameters to output weld position with standard signal or wireless network communication. It is a general-purpose vision sensor for the field of seam tracking.

In order to achieve the above purpose, the technical solution of the present invention is to provide a welding seam tracking visual sensing device based on laser structured light, which can collect and process weld seam feature images, and output weld seam position signals to control welding robots or control The welding process is carried out with the welding torch on the device; a) the laser beam emitted by the inline laser is used to generate the laser structured light plane to form structured light stripes; b) the camera equipped with a filter lens is used to collect the laser stripe image of the weld seam characteristics; c) the laser The structured light plane forms an angle of 30°~60° with the central axis of the camera head.

The sensing device includes a laser, a camera, a welding torch and a processing unit, wherein the laser and the camera are packaged together, and the welding torch is rigidly connected to a frame, the camera has an optical filter, and the sensor and the welding torch are fixed, The direction of the central axis of the two is the same, and they are in the same plane. The plane intersects the plane formed by the welding seam and is perpendicular to the surface of the workpiece to be welded; the camera is electrically connected to the processing unit, and the processing unit is electrically connected to the welding robot or controller through the signal line .

Described sensing device, its described processing unit comprises computer and its parts, control panel, monitor and signal wire, and computer is electrically connected with control panel, monitor and signal wire respectively, and video camera is electrically connected with the image acquisition card of computer. connection, the processing unit is also provided with an analog output interface and a wireless network interface.

In the sensing device, the camera is an industrial camera.

In the sensing device, the laser structured light plane and the camera are calibrated, and the offset of the welding seam is calculated by the computer from the coordinates of the characteristic points of the structured light stripes.

Said sensing device, said analog quantity output interface and network interface, can output the welding seam position offset information as a standard analog signal or read it out through a wireless network.

The sensing device can be set and adjusted according to the type of welding seam by using the network or selecting through the buttons on the control panel.

In the sensing device, the computer is also electrically connected to the weld image processing module. The weld image processing module has data of various types of welds, is used to extract the coordinates of weld feature points, and has extended functions.

In the sensing device, the optical filter can transmit light waves with a wavelength of 670nm.

The signal output of the sensing device has horizontal and vertical weld position deviation outputs.

The outstanding feature of the present invention is that a line laser is used to generate a structured light plane, a CCD camera equipped with a filter lens is used for image collection, and the welding seam position information is obtained through image processing.

The technical solution adopted by the present invention to solve the technical problem is: the sensor takes the PC104 computer as the center. A beam of line light source emitted by a straight-line laser constitutes a structured light plane, which is projected onto the surface of the workpiece to form structured light stripes. After the CCD camera receives the reflected light, the video signal is sent to the image acquisition card, which is converted into a digital signal by A/D. According to the weld type and process selected on the sensor control panel, or the system parameters set through the wireless network, call the corresponding image processing program (image filtering, enhancement, extracting the center line of the weld, extracting the feature points of the weld) to obtain the weld After the seam deviation signal, the control signal is obtained through conversion, and the welding seam position offset information can be output by standard analog signal (1~5V, 4~20mA) or wireless network card communication output as the actuator through the analog output interface and network interface. Provide control signal output to achieve the purpose of real-time deviation adjustment.

The beneficial effect of the present invention is that it can be adapted to various types of welding seams and processes, provide welding seam position signals, and is a universal, flexible and independent welding seam tracking vision sensor.

Description of drawings:

Fig. 1 is a schematic diagram of a laser structured light weld seam tracking visual sensing device of the present invention;

Fig. 2 is a schematic block diagram of the welding seam tracking visual sensor based on laser structured light in the present invention;

Fig. 3 is laser images of different types of welds collected by the sensor of the present invention.

Detailed ways:

See Figure 1. In Fig. 1, a semiconductor laser (LD650-5-3) 1, a WATEC505 black and white industrial camera 3, an optical filter 2 and a splash guard 5 are packaged in the vision sensor housing. There is a hole on the top of the sensor housing for the power and signal lines of the camera and laser to lead out. The sensor is electrically connected to the processing unit 7 through a video signal line. The bottom of the sensor casing is empty, and the side where the camera is installed is covered by a splash guard 5 made of plexiglass, which can block welding spatter and protect the camera. The side of the sensor casing has a hole, and is rigidly connected with the welding torch 4 through a steel beam. The camera in the sensor case is installed vertically and fixed through the hole on the back of the sensor case; the semiconductor inline laser 1 is obliquely fixed in the sensor case by the mounting bracket, and the position and angle of the laser 1 can be adjusted through the mounting bracket. Adjust the angle between the laser 1 and the camera 3 so that the center line of the optical axis of the camera 3 forms an angle of 30°~60° with the laser structured light surface, which is 45° in this example. Adjust the center line of the optical axis of the camera 3 to be perpendicular to the surface of the welded workpiece where the weld seam 6 is located, and make the height of the bottom of the sensor casing from the surface of the welded workpiece to be 10-20 cm. The laser beam emitted by the semiconductor inline laser 1 generates a structured light plane, which irradiates the surface of the welding workpiece to form the characteristic laser stripes of the weld 6 . The optical filter 2 below the black-and-white industrial camera 3 can pass through light waves with a wavelength of 670nm. The industrial camera 3 collects the laser stripe image formed by the inline laser 1 . The laser line width is 0.5mm, the spread angle is 20°, and the power is 5mw.

The image of the weld seam 7 is sent to the processing unit 7 as a standard PAL signal. Through the image acquisition card (micro-view MOKA-C10), it is converted into digital signal through A/D. Processing unit 7 adopts PC104 computer and other components, including image acquisition card, PC104 bus switch input/output board PM-582, PC104 bus analog output card PM-520, wireless network card LucentPC24E-H-FC. PC104 computer and its components, together with a control panel 8 and a monitor 9 constitute a processing unit. The camera 3 collects the laser image with the characteristics of the weld 6, and the processing unit 7 invokes the weld image processing module to extract the image coordinates of the weld stripe feature points.

Since the laser plane and the camera 3 have been precisely calibrated, the conversion matrix from the two-dimensional space of the image to the actual three-dimensional space can be determined through calculation, so that the actual welding can be calculated from the two-dimensional coordinates (u, v) of the weld stripe feature points in the image space. The three-dimensional coordinates (x, y, z) of the sewing point. For the precise calibration method of the laser plane and camera 3, please refer to the patent document (Application No.: 02158341.2 Name of Invention: Laser Structured Light Vision Sensing Device). And because when the welding torch 4 moves left and right in the horizontal direction, the weld stripe image also moves left and right; when the welding torch 4 moves up and down in the vertical direction, the weld stripe image also moves up and down, so the weld seam 6 on the image The horizontal offset and the vertical offset of the feature points respectively represent the corresponding offsets of the welding torch 4 relative to the weld seam 6 . In this way, the horizontal and vertical deviations between the feature point coordinates of the weld image and the standard position can also be used to represent the horizontal deviation or vertical deviation between the welding torch 4 and the standard weld position. Therefore, the visual sensor can output the three-dimensional coordinates of the actual welding spot, or take the horizontal and vertical deviations between the welding torch 4 and the standard welding seam position as the output.

The processing result of the sensor (three-dimensional coordinates of the welding spot or the horizontal deviation or vertical deviation signal between the welding torch 4 and the weld 6) can output a standard analog signal (1~5V, 4~20mA) through the signal line 14, or can be read out through the wireless network . Two welding seam type BCD encoders 11, each encoder can select ten states from 0 to 9, a total of 100 states in combination, can choose to set a variety of welding seam types; "process selection switch" 12 selects once/multiple Secondary welding seam; there are indicator lights 10 and control buttons 13 on the control panel 8; operation, normal and fault indicator lights can indicate the status of the sensor; the "start" button determines the starting position of the weld seam as the standard position, and the sensor starts to work; "stop" The button is used as a welding stop sign, and the sensor stops outputting. The function setting of sensor and the input/output of state can be realized in two ways: the control button 13 on the control panel 8 and indicator light 10 can set the parameter (welding seam type, procedure) of sensor, control the working project of sensor ( Start, stop, reset), and indicate the working status of the sensor (working, stop, fault); it is also possible to set the parameters and control information of the sensor through wireless communication, and read the status of the sensor.

The working process of the sensor can refer to the functional block diagram of the sensor in Figure 2. The weld seam laser image collected by the camera 3 is sent to the monitor 9 for output, for the operator to observe the weld seam image in real time; and the weld seam image is sent to the PC104 processing unit 7 for image processing. Through the selection command of the control panel 87 or the parameters set by the wireless network, the type of the weld 6 is determined, and the corresponding program module in the weld image processing library is called to process the weld image. In the weld image processing program library, specific image processing modules (such as V-shaped, trapezoidal, butt joint, lap weld image processing, etc.) have been developed for the image characteristics of various weld types, forming a weld image processing function module library , and can add a new weld type image processing module with extended functions. Referring to Figure 3, the figures are (a) lap type welds, (b) trapezoidal secondary welds, (c) V-shaped butt primary welds and (d) direct butt welds (with a gap of 2 to 3 milliseconds in the middle ) laser image. By calling the weld seam image processing module, the image coordinates of the weld seam feature points are extracted, and then output by an analog signal or by a wireless network. In this way, the weld seam tracking visual sensor can collect and process the laser image of the weld seam and output the position signal function to control the welding robot or the welding torch 4 on the controller to carry out the welding process.

Claims (5)

1, a kind of weld joint tracking vision sensing equipment based on laser structure light can be gathered, the processing welding lines characteristic image, and output position while welding signal, carries out welding procedure with the welding gun of controlling on welding robot or the controller; It is characterized in that, comprise sensor casing, laser instrument, video camera, welding gun and processing unit, wherein, the power line of video camera, laser instrument, holding wire are drawn by sensor casing top drilling; The sensor bottom of shell has transparent splash guard in the side that video camera is installed, and sensor casing side is rigidly connected by girder steel and welding gun, and video camera vertically is installed on sidewall in the sensor casing, and the camera photography headband has optical filter; The laser instrument mounting bracket tilts to be fixed on another sidewall in the sensor casing; The camera light shaft centre line becomes 30 ° ~ 60 ° angles with laser structure light face, the camera light shaft centre line is vertical with weld seam place welded piece surface, and sensor casing baseplane is 10～20cm apart from the welded piece apparent height; Video camera is electrically connected with processing unit, and processing unit is electrically connected with welding robot or controller by holding wire.

2, sensing device according to claim 1, it is characterized in that, described processing unit, comprise computer and parts thereof, control panel, monitor and holding wire, computer is electrically connected with control panel, monitor and holding wire respectively, video camera is electrically connected with the image pick-up card of computer, and processing unit also is provided with analog output interface circuit and radio network interface.

3, sensing device according to claim 1 is characterized in that, described laser instrument is semiconductor one word laser instrument, and video camera is industrial camera.

4, sensing device according to claim 1 is characterized in that, described optical filter can be through the light wave of wavelength 670nm.

5, sensing device according to claim 2 is characterized in that, described computer also is electrically connected with the weld image processing module, and the weld image processing module is equipped with the data of multiple weld seam type in advance.

Images