Vision-based checking and image processing for melt pool of laser deep penetration welding

On the basis of analyzing the infrared spectrum characteristics of weld pool by Nd:YAG laser,the co-axial visual monitoring system for Nd:YAG laser deep penetration welding of stainless steel was constructed successfully,by designing the transmitting and separating structure for weld pool radiation and YAG laser,and then optimizing the suitable light components, such as half penetrating and half reflecting lens,narrow band pass filter,neutral weakener,variable-focus lens and so on,according to the requirements of co-axial monitoring. By means of the developed monitoring system,the clear images of weld pool were attained. In allusion to the recorded image of weld pool, gauss filter,edge enhancing function and algorithm for edge tracking were developed,by which the starting point and rules for edge tracking were put forward. Consequently,the edge of Nd:YAG laser welding pool was extracted from CCD images, which would be helpful to the on-line monitoring and controlling of laser deep penetration welding for better welding quality in future.