Algorithm Research On Image Recognition About Welding Seam Of Buick Automobile Auxiliary Frame

With the rapid development of modern industry such as metallurgy, machinery,petrifaction, the atom, spaceflight and so on, the welding industry becomes more and more important. It has become the third critical kind of fabrication industry, only less important than assembly and machinery process, so the assurance of the welding quality is of big importance. The quality of the welding seam includes shape, position and fault. Especially in the auto industry, it is important to detect and estimate the welding seam.The content of this thesis is one of the key technologies in the Welding Seam Recognition System, which is a project of Shanghai Auto Fund. We developed this system to detect the quality of the 21 welding seams in Buick automobile auxiliary frame. Because the position and the shape of the 21 welding seams are different, a number of CCD cameras are used to get images. All these cameras communicate with computer via USB interface. Then, we process the captured images and estimate the quality of Welding seam.Welding seam image of Buick automobile auxiliary frame is very unstable.Because it is welded by human, and it is got by CCD. The detect module based on median filtering and KL transform is adopted to detect welding seam image. But the question of accuracy and speed is still not sloved.This thesis focuses on the research into the question of accuracy and...