| With the rapid development of manufacturing industry in China, improving the welding efficiency and welding quality is the key to make it faster and better. Using robotic automatic welding technology provides a perfect solution for this requirement. Welding automation not only improves welding efficiency and welding quality, but also enhances the working conditions and improves the working environment of welders.Today, the commercialization of the welding robot is used to teach programming and then do the welding work. This way can’t make judgments on the welding track deviation result from welding thermal stress, tooling error, machining error, etc. In this paper, the PC+FPGA based robot automatic welding system can do the automatic weld-seam tracking without repeated teaching programming. The PC terminal control software mixed programming methods, and programmed by Open CV and MFC. It integrates the interface and image processing function, which not only do the human computer interaction and welding seam image recognition, but also do the automatic tracking and send motion control command. FPGA controller is the link and bridge of the controlling system, which can receive and process the signal, complete the serial protocol generation and decoding work. The software and hardware of FPGA controller is designed by Verilog HDL and PROTEL99 SE. In the design, ABB robot controlling program regulates the control process of the control system, sends data to the the host computer to display the current status information.In this paper, we use composite sensing technology. The linear displacement sensor and CCD image sensor can detect CTWD and lateral position of the weld seam. The PC terminal control software uses the data and image processing to identify the change of the height and lateral position of the weld seam. At the seam time, the PC terminal control software sends the deviation control command and finally finishes the seam tracking function.By using camera calibration experiment and the calibration experiment of position relationship between the camera and the torch, the transformation relationship between the target point in the world coordinate system and the pixel coordinate system. We put forward a new calibration method to capture the relationship between the camera and the torch, and the error and be controlled at the sub pixel level.Through different ways of welding seam tracking experiments, we obtain the influence of the curvature change of the welding seam to the seam tracking accuracy. And by the experimental verification, when using the welding system, the tracking accuracy of CTWD is about ±0.4 mm, the lateral tracking accuracy is about ±0.7 mm, which meets the practical application requirements. |