Identification Method of Welding Seam at Riveting Point Based on Visual Sensing
In order to solve the serious interference problem of welding seam identification caused by riveted welding points in automatic welding of assembly parts,a method of weld identification based on visual sensing is proposed.First,a CCD industrial camera is used to collect the weld image with laser strip,then the median filter is used to smooth the noise,the images are segmented with OTSU adaptive thresholding,the laser stripes are refined with skeleton extraction,a straight line is fitted with probabilistic Hough transform,and a binary matrix based method is used to remove the interference of riveted weld joints on the recognition of weld seam in the ROI region.Finally,the center point of weld is found by using the endpoint coordinates obtained by line fitting.The experimental results show that the error of this method is less than 0.6 mm,which can quickly extract the weld feature points and meet the welding requirements.