Online Weld Quality Detection In TIG Welding Based On Visual Information

The TIG welding is one of the common technologies in boiler welding. At present, real-time analyzing and detecting by human and the pressure testing after welding are two main kinds of methods for boiler welding inspection. The problems of these methods are labor intensive fatigable, subjective, inconsistent, high-cost and slow production speed. So non-destructive detection of weld defects has been concerned, and the detection of weld defects based on machine vision is a hot topic in recent research at home and abroad.The technology of online weld quality detection in TIG welding based on visual information was systematically studied. The research works of the paper are as follows:1. The system of online weld quality detection in TIG welding was developed, the comparatively clear images of weld pool area were obtained and the online weld quality detection was completed using this system.2. The welding arc and the weld pool were identified, and their geometric parameters were also quantified and extracted through the research on their image processing, because the shape of the welding arc and weld pool have obvious response to the welding defects.3. The influence of the weld quality in different welding parameters was studied, and the difference of the welding arc and weld pool geometry parameters with different welding quality were analyzed, the characteristic parameters which could be used to estimate the welding quality have been found, which provided a theoretical basis for the detection of weld quality. The classification problem of different welding quality was solved using these characteristic parameters in combination with Support Vector Machine(SVM). Compared to the other classifier, the results of SVM algorithm were the best, and the time required was less. The results prove that SVM algorithm is an effective method for the classification of the weld quality. So this method can be used for the online weld quality detection, to reduce labor costs and improve work efficiency.