Visual measurement and tracking in laser hybrid welding

This paper presents a novel system for the automatic analysis of a hybrid welding process. High-speed imaging and laser illumination are used to measure the regularity of electric arc frequency and flight directions of filler metal droplets. A fuzzy c-means clustering method is used to detect arcs and segment the video sequences. The droplets are localized by combining principal component analysis and a support vector machine classifier. The flight of a droplet is tracked using Kalman filtering. Experiments indicate that the system is able to track the flights of droplets and to determine the regularity of the arc frequency with a high accuracy if the imaging conditions are stable.