Study on Automatic Recognition Technology for Exhaust Pipe Weld

Auto industry is the highest degree of the industrial automation, such as assembly, painting, fixed spot welding, but the exhaust pipe parts still have been welded by hands, the welded quality and its costs are difficult to be controlled besides its technology and the work environments. study on identification and location of the exhaust welded pipe with image processing to achieve the robot welding is work in this paper. System adopted the CMOS image sensor OV6620, with FPGA Module designed to achieve high speed data acquisition, analysis and location of weld seam image is completed by the NIOSII. The system is simulated with the involved algorithm by Mat lab, and gets the weld information based on the characterize of the maximum Connected-component in noise environment, with Canny algorithm for recognition of weld edges information, the center coordinates and weld seam width information are got last.