Research Of Weld Location Detection System Based On Image Sensing Technology

The weld image processing is an important part of automatic seam tracking,the key issues of today’s research is to improve the welding quality and reduce the labor intensity of welding workers,so the study of welding seam image processing algorithm is of great significance.Study on the related aspects of reference at home and abroad,combined with the actual needs to weld image processing based on laser vision as the research object,focusing on all aspects of the problem of weld in image processing methods were carried out.The main work of this paper is as follows:(1)The image contains a large amount of information and other issues,We take the lower brightness,set the ROI image acquisition area,make gray processing method of image to get complete information and little noise.The result shows that these methods made the acquisition of high quality and effective.(2)Aiming at the problem of noise acquisition process,We take the median filtering method which can effectively remove noise and noise like reflected light in weld image.Remove the noise while the maximum protection of the edge information of the image.According to the weld image after filtering,the weld image binarization threshold are selected automatically by Otsu algorithm.The results show that this method in image processing is stable and effective.(3)According to the problem that weld image edge and center line extraction are not accurate,using Canny operator edge detection algorithm and Zhang-Suen thinning algorithm based on center line extraction method.The treatment effect compared with the Roberts operator,Sobel operator,Prewitt operator,Laplace operator in edge detection shows that the edge detection based on Canny operator can be better.In the extraction of center line through the comparison based on Rosenfeld thinning algorithm,based on index table image thinning algorithm and pixel based image thinning algorithm over the treatment effect of calendar,indicates that the center line extraction method based on Zhang-Suen thinning algorithm is stable and effective.(4)According to the angle of the influence of weld seam feature points extraction problem,the paper proposes to extract the pixel coordinates of feature points by using the line fitting intersecting method based on Hough transformation,in this paper the performance of pixel corner detection method and method of extraction times calendar pixel coordinates of feature points were compared,the result shows that the line fitting intersecting method based on Hough transformation is effective,the algorithm is simple and easy to realize.(5)The weld location detection system is built,and the visualization software of the weld location detection is completed.The research results show that the method proposed in this paper can effectively extract the position coordinates of T joint angle weld spot.