Research on Quality Control of Arc Welding Robot Based on Molten Pool Contour extraction

Abstract

In order to overcome the influence of various uncertainties on the welding quality in robot welding process, this paper proposes a multi-information fusion detection method. At first, in order to obtain the geometrical characteristics of the molten pool image, the adaptive Wiener filtering algorithm was used to denoise the image, and the Otsu threshold method was used to binarize the image. Then the binarized image was co rroded, swelled, removed and small area operation. The complete clear edge information of the weld pool is extracted; then the simple and effective geometric feature parameters are used to describe the shape of the weld pool. According to the characteristics of the arc welding temperature field, the method of radiation temperature measurement and the wavelength selection principle are discussed. From the arc temperature, wavelength and the angle of the thermometer with respect to the molten pool area, the interference of the arc at the temperature measurement point in the near-melting pool area was analyzed; It is proved quantitatively that the radiation temperature measurement method is feasible under the condition of small arc current.