Research On Robot Welding Feature Recognition And Welding Seam Correction Of Steel Structure Parts

In the welding process of large steel members,the intelligent welding robot not only has higher operating stability than the manual and higher machining accuracy than the visual measurement,but also can improve the machining efficiency and reduce the production cost.Therefore,the research of robot intelligent welding technology is paid more and more attention by the industry.With the continuous improvement of off-line programming,visual sensing,image recognition and processing and robot welding correction technologies,the above achievements have the possibility of joint application in the field of robot intelligent welding.As one of the indispensable technologies for intelligent robot work,offline programming can improve the adaptability of welding robot in the actual production environment.Among them,DXF format files of CAD drawings are converted into off-line programming mode of robot identifiable program,which can reduce the inconvenience brought by on-site teaching.Moreover,if the visual sensing equipment is used to correct the welding seam error,the higher welding success rate can be obtained,which makes the robot welding more in line with the actual production needs.Therefore,this paper will design a robot welding system for welding and welding seam correction of steel structural parts.The specific research contents are as follows:(1)in order to improve the flexibility of industrial robot welding and avoid frequent manual on-site teaching,this paper studied the selection and extraction of DXF file information through C++ programming method,and analyzed the welding feasibility of robot shaped steel components;A human-computer interface is designed which can clearly reflect the welding coordinate and image information.The complex parts are calibrated by six-point calibration method and kinematic correlation formula of robot.(2)to protect welding work conform to the requirements of the processing precision,calibration work should be carried out on the visual sensing system,determine the camera installation,to structure light sensor calibration,parameters optimization for the camera for a more precise,complete study of the principle of the camera calibration,the calibration method of MATLAB with analysis,and carries on the improvement of contrast,and through the camera calibration experiments to verify this theory data analysis.(3)in order to make the weld image clearer and easier for observation and subsequent processing,it is necessary to carry out image preprocessing,analyze and study the types of noise and filter,and remove noise interference.After the noise reduction,the image is processed into grayscale,and other interference factors are excluded.After image segmentation,the clear image of the weld is obtained,so as to make preparations for the next step of welding seam correction.(4)off-line welding technology is commonly used in the welding of steel structural parts because of its efficiency and efficiency.However,due to non-linearity,errors often occur and it cannot be automatically adjusted.Therefore,off-line welding seam correction technology is adopted to improve the welding effect of the robot.Aiming at the hardware architecture and operation flow of the welding robot structural light vision correction system,and based on the KUKA welding robot,a correction algorithm for the starting point and the changing point is proposed,and the simulation experiment is carried out by MATLAB.Finally,the robot welding and calibration system for steel structure parts is set up.According to DXF off-line programming technology,the software of DXF file information screening and extraction is designed and tested.According to the test results,through C++ programming,we can obtain the refined information of the graph elements and paths required by robot offline welding programming.Applying DXF welding to robot welding technology can effectively reduce the frequency of manual instruction by operators.With the help of the research on the calibration method of the visual sensing equipment,the installation and debugging of the image acquisition equipment are completed.Two different correction algorithms were used to complete the establishment of the correction system.Finally,the experimental results show that the research of robot welding feature recognition and welding seam correction technology for steel structural parts is feasible,and the method presented in this paper has a great improvement in efficiency and accuracy compared with the traditional method.