Seam tracking algorithm based on magneto-optical imaging and self-adaptive Kalmanfiltering

Accurate seam tracking is a prerequisite for laser welding with good quality. A seam tracking method based on Kalman filtering with colored noises is proposed to predict the seam deviationsin micro butt joint whose width is less than 0.05 mm. In the experiment, the weldments were magnetized by usingan excitation magnetic field. Meanwhile, a magneto-optical sensor based on the principle of Faraday magneto effect was applied to acquire the magneto-optical image of the weld joint. By analyzing the magneto-optical images of weld joint, the joint center position was extracted and defined as the state vector. Then the state equation and the measurement equation based on the weld joint center position were established. Considering that the system process noise was colored noise, the Sage adaptive filtering was used to lessen the noise influence.The innovation series was used to estimate the process noise variance matrix, and the weld joint position could be predicted accurately. Experimental results show that seam tracking accuracy can be improved effectively with self-adaptive Kalman filtering method.