Abstract: The application of Kalman filtering for the weld detection in the arc welding process is presented. A Kalman filter is applied for processing the arc weld pool images from a visual sensor to recursively compute the solution to the weld position equations which are established based on an estimation of the position and displacement of the centroid of the weld pool images. This centroid whose characteristic corresponds with the weld position is extracted as the weld measurement eigenvector. The evolution of the weld position data from the weld pool images can be described through a state equation and a measurement equation of the image centroid. Based on a color noise model, the advantages of Kalman filtering over least squares approaches is taken to estimate the weld feature location on the image plane at the next sampling time and reduce the weld measurement error caused by the process and sensor noises. Computer simulations and actual welding experiments are demonstrated the effectiveness of the proposed algorithm in the presence of weld pool image noise and are tested the robustness of weld position detection.

Key words: Kalman filter, State estimation, Weld detection, Weld pool image