Research Of Image Technique In Tracking Processing

Welding plays an important role in metallic material processing. Welding automation can improve operators of welding quality and productivity greatly. It will also relieve the human tedious and repetitive welding. Seam tracking is an essential technique to fulfill welding automation. With rapid development of computer science and image processing technology, vision seam tracking becomes a research hotspot at home and abroad recently.In context of the assembly line construction project of square cabin of the large modes' 13m line welding, the whole procedure of laser seam tracking is investigated and a seam tracking system is devised. It includes vision sensing system, characteristic quantity detection and deviation recognition system, fuzzy control system and execution system.An module of line light method is used in this paper... As an added light source, laser is projected in front of seam and the light stripe reflecting from the work-piece is received by CCD Camera. After being quantitated, the digital image is saved in the main computer memeory. On the basis of image processing using LOG filtering, binarization and median filtering, Roberts edge detection, Hough transformation characteristic quantity detection, deviation recognition using slope method are implemented sequentially. The deviation is recognized finally. Before control, deviation must be filtered. Only those between control range can be controlled. Without using conventional fuzzy control, the parameter self-tuning fuzzy control which is widely applied in the numerical control field is introduced. This conquers the weakness of unchangeable quantitative factors in conventional fuzzy control. In order to fulfill real time seam tracking, two dimensional input deviation is fuzzy controlled and control instructions are sent to alternating current servo system.The whole time is less than 200ms, thus five images can be captured within one minute. This can meet the demand of real time seam tracking.