Research On Weld Detection And Grinding System Based On Line Laser Sensor

Traditional welding seam grinding is mostly done by hand with grinding tools,which is inefficient and difficult to meet the requirements of stability and accuracy.The rapid development of vision technology and sensor technology has provided new ideas and methods for seam detection and robotic grinding automation.Laser has the characteristics of strong monochrome and high brightness.Compared with CCD camera,it has stronger anti-interference ability.Therefore,this paper carries out the research of seam detection using line laser sensor.The seam wheel is collected by sensor mounted at the end of the robot.The outline data are processed to lay the foundation for robot grinding.Firstly,the composition and working principle of the whole system are introduced.A hand-eye calibration method based on standard sphere center is proposed,and a more accurate hand-eye relationship matrix is obtained.Secondly,the data analysis and processing methods,including data denoising,data smoothing and data reduction,are studied.They are based on distance statistical analysis method,bilateral filtering method and point-to-point normal vector reduction method,respectively.On the premise of data characteristics,noise interference and data processing amount are reduced to a great extent.Then,a seam feature point detection algorithm based on maximum distance analysis is proposed,and the residual height and fusion width are calculated according to the feature points.Finally,the advantages and disadvantages of least square method,cubic spline interpolation method and cubic B-spline fitting method are compared,and the cubic Bspline method is chosen to fit the curve.Spatial weld path fitting method of strip is used to get the grinding posture of robot.In order to realize data visualization and hardware interactive communication,a seam detection software based on open source algorithm library PCL and Qt was developed,and the feasibility and accuracy of the system were proved by seam detection and comparison experiments.The grinding experiment of guided robot was carried out to verify the processing situation based on the curve fitting of seam path.