Research On The Control Method Of Welding Seam Image Recognition And Tracking

Welding automation sees an increasingly range of applications in the industrial field,especially in some welding processes of large steel structure equipment. Manual welding cannot ensure the quality and the efficiency of welding because of the affect of human factors.Besides, some special position can not be welded by hand, thus some other automation weldingmethods are urgently needed to replace traditional Manual welding. Welding automation hasthe features of high efficiency and reliable welding quality, which is more and more popular andis widely used in industrial fields. With the advancement of technology, the main developmenttrend of welding automation is the seam tracking control system based on vision sensors. So ithas great significance for welding work to do research on seam recognition and tracking control.The main content of this paper is as follows:(1)Research on welding seam recognition algorithm. Firstly, through the CMOS camerato collect data, and then fed it into the field programmable gate array (FPGA),where the imagerecognition algorithms of median filtering, binarization, and Hough change are carried. Noisesof welding images are fast removed from their complexity textures, and the features informationof the welding images is extracted. Parallel processing is implemented in the algorithm withinFPGA, which greatly improve the real-time nature of the welding seam image recognition.(2) Build the welding seam image recognition hardware circuit and do the softwaredesign on it. Use EP2C35F484C8as the main chip, related peripheral circuits are designed. Dothe top-level module division on the entire system, then hardware description language (Verilog)are edited, besides, simulated on each module, finally, downloaded and debug the validationprogram. Compared with other welding seam recognition systems, The FPGA welding seamidentification system has the characteristics of faster computing speed, and higher real-time andsignificantly features extraction ability.(3) According to the kinematical model of the welding robot, we designed a Fuzzy PIDswitch controller. To select a control mode according to the value range of the welding seam deviation, and switch different controller according to input values, which enables rapidtracking. Then according to the model of the design, Matlab/Simulink tools are used to dowelding seam tracking module simulation. After comparing the simulation results of PIDcontroller and Fuzzy PID switch controller, the study shows Fuzzy PID switch controller hasbetter real-time and anti-jamming capability than PID controller.