Image processing of seam tracking system using laser vision

A method of real time image processing of laser vision sensing seam tracking system has been studied, an algorithm of single point filtering and a binariantion processing by which the severable value is scoured automatically are put forward, getting rid of the noises effectively. A feasible method has been proposed for the laser midline extraction, by adopting inside stalk transformation, a single and consecutive laser midline can be distilled. Finally the character point is detected by the method of analyzing slop. It could detect the character point exactly and rapidly, and meet the demands of the real time seam tracking system.