Image Processing of Aluminum Alloy Weld Pool for Robotic VPPAW Based on Visual Sensing

The intelligent welding technology is one of key development directions in “Industry 4.0” systems, and the use of sensing method is the main research direction of future intelligent welding technology. Due to the characteristics of aluminum alloy variable polarity plasma arc welding, by employing a complementary metal oxide semiconductor welding camera with a composite filter glass, a clear image of weld pool can be captured by side vision sensor. Aiming at the images captured with digital graphic processing technique, in this paper images are processed by Labview vision module. Moreover, in this paper, a method for image processing, which includes reversing, local threshold determination based on variance within the group, fast Fourier transform low-pass truncate filtering and advanced morphological operations for particle, is proposed to extract the characteristics of weld pool width. The result has shown that proposed image processing system can obtain reliable information on width of weld pool, which represents the basis for realization of quality control in aluminum alloy variable polarity plasma arc welding.