Research On Recognition Of Filmy Board Lap Welding Seam Position Based On Surface Laser Vision Sensing

With the rapid development of industrial automation,automatic welding has been widely used.Lap welding of filmy board has a great demand in real industry,but at present,lap welding is mainly manual welding or teaching welding,the cost of manual welding is high,the environment is bad,and the welding efficiency is low.The characteristic information recognition of filmy board lap welded seam is not obvious by line laser,only a few feature points are used to fit the weld seam,error is large.Aiming at the deficiency of extracting lap weld of filmy board with line laser vision sensor,the surface structured laser vision sensor is designed by using the surface structured laser of wavelength 650 nm as the auxiliary light.When the surface structured laser is tilted at a certain angle,it will form a weld shadow on the workpiece,the surface structured laser contains a characteristics of filmy board weld.It overcomes the difficulty of large error of fitting weld line caused by traditional line laser only through few characteristic points to determine the lap welding seam of filmy board.Lens installation will distort the image.This paper introduces the reason of imaging distortion of CCD and corrects the distorted image by Zhang Zhengyou correction method.Surface structured laser vision sensor and DH-CG410 image acquisition card are used to collect the image of lap welded seam of filmy board,and welding seam is extracted from the lap welded seam image of filmy board.Pre-processing of lap weld seam extraction of filmy board includes surface structured laser ROI extraction,adaptive mean filtering for noise reduction,image segmentation,solitary point filtering,closed operation,edge extraction.Finally,the strip lap weld is extracted by column search combined with the distance between two points.The height recognition model is established by using the geometric position relationship between CCD and surface structured laser.The parameter calibration andheight recognition experiments of the established height recognition model are carried out.The error of the height recognition model is within 0.35 mm.Combining with the height recognition model,the initial welding point recognition of lap welding seam of filmy board based on surface structured laser vision sensing is studied.The initial position error of the lap welding seam of filmy board is less than 0.3mm.The deviation identification model of lap welded seam of filmy board is established,and the deviation of weld seam detected is less than 0.37 mm.Based on VC 6.0 and OpenCV1.0 software,the information extraction interface of lap welded seam of thin plate is designed.The interval of image acquisition is40 ms,and the main frequency is 3.20 GHz,memory is 2G industrial control computer is used to process the image of lap welded seam of thin plate.It takes less than 25 ms to process an image,which meets the real-time requirements of welding.