| With the continuous improvement of industrial production technology and its degree of automation,the welding technology of steel structural parts requires higher welding quality,and the traditional manual welding technology can no longer meet the needs of actual production.Most steel structure parts have the disadvantages of many varieties,small batches,poor consistency of weld shape and position,and complex robot repeated positioning process.It is difficult to realize automatic welding.Using welding robots to replace traditional manual welding can improve welding efficiency and welding accuracy.Aiming at the problems existing in the welding of steel structure parts,this paper applies laser vision to the welding robot,and studies the intelligent identification of the welding seam of the steel structure welding robot based on laser vision,so as to complete the intelligent identification task of the welding seam of the steel structure parts.Firstly,the camera pinhole imaging model and camera distortion model are established,and the calibration of the vision system is realized by Zhang’s calibration method,the internal and external parameters of the camera are solved,and the distortion coefficient is calculated considering the camera lens distortion.The hand-eye calibration method is studied to solve the hand-eye relationship matrix,and the relationship between the pixel coordinate system of the image and the world coordinate system is obtained.The relative positions of the industrial camera and the laser in the laser vision system are designed.The industrial camera is used to irradiate the surface of the steel structure vertically,and the laser is irradiated obliquely at an angle of 45 degrees.In order to extract the position coordinates of the feature points of the steel structure welds more accurately,a set of intelligent identification algorithms for the welds of steel structures is designed,and the feature information of the weld images of the steel structures is extracted according to the established welding process.Including image grayscale,power-law transformation nonlinear grayscale processing,median filter denoising,image threshold segmentation and other preprocessing operations,the weld features with laser strips are extracted.Aiming at the obvious irrelevant connected areas and blobs in the image except for the laser stripes,an adaptive Otsu threshold method combined with the morphological opening operation is proposed to process the weld image and remove the interference information perfectly.The detection effect and processing time of the five edge detection methods were compared,and the Canny operator with the best detection effect was selected.The geometric center method was used to extract the center line of the laser strip,and the Hough transform was used to fit the center line of the laser strip.Process to obtain the position information of weld feature points.Finally,an experimental platform for intelligent identification of welding seam of steel structure welding robot based on laser vision is built.The software uses the Visual studio platform to complete the software composition and control interface,conduct intelligent identification experiments of welding seam of steel structure parts,and compare the theory of any seven groups of welding seams.According to the welding seam identification result,it can be seen that the error can be controlled within the range of0.87 mm,and the identification accuracy can meet the actual welding needs. |
