| Welding is one of the important techniques in manufacturing. Driven by many factors, automation and intelligence of welding have become the trend of welding industry. Weld seam tracking technique plays a great part in the field of automation and intelligence of welding. Weld seam tracking technique can not only improve the stability of weld quality but also make the operation safer and more economical. The advanced weld seam tracking technique based on image processing has been widely concerned in the whole world.In order to supervise the weld seam pool, a software system has to be programmed to analyze the image captured by CCD visual sensor. The specialized software system which can be transplanted easily was programmed by language VC6.0. It can control the hardware directly and lay good basis for later close-loop control.The system mainly includes: point operation, image transformation, image enhancement, image restoration and image analysis. The median filter with the size of 5×5 can filter some bigger burr and reduce the operating time by one and a half time. Confirmed by calculation, the time and space complexity of this system can meet the real time manipulation of high-speed laser welding. The enhanced gradient operator was designed to stand out the edge of image and retain the low frequency information.Aiming at the realization of weld seam image processing, a hardware- structured system for the weld seam auto-tracking image processing has been worked out. Based on the system, the arithmetic of the image processing was analyzed systematically and the process flow as well as the arithmetic of the structured light-based weld seam Image processing was compiled. The arithmetic realized the feature extraction of the weld seam Image. As been used to control the Weld seam auto-tracking process, it can reduce the labour force, improve the... |