Research Of The Automatic Welding Compound Vision Guide System

In recent years,robot welding has replaced artificial welding in more and more areas,and robot welding has gradually developed toward the direction of intelligence with perception.In the development process,the use of sensors to obtain the welding parameters has become the mainstream of research on automated welding,which determines the accuracy and quality of welding.The automated welding composite vision guidance system studied in this paper includes vision sensor module,upper computer control module,guide rail movement mechanism and lower machine control module.The vision sensor module also includes a bi-color LED strip light source,a laser structured light,and a CCD industrial camera.This paper analyzes the current situation of automated intelligent welding research at home and abroad and the application of machine vision technology in the field of welding,and thoroughly studies the key technologies that need to be solved for automatic welding systems with standardized shapes and sizes of standardized workpieces.The welding of ribs,connectors,stampings,castings,etc.is a type of welding problem often encountered in welded structures.These welding workpieces are often characterized by small size,consistent shape,and some geometric features.The light source is illuminated on both sides of the CCD camera and taken for image processing to obtain the plane coordinates of the welding workpiece on the X and Y axes.Through the cooperative operation of the upper and lower position machines,the welding torch and the sensor are moved to the center position of the welding workpiece obtained.Open the laser structure light,collect the image and process it to obtain the depth information of the weld seam.The lower position machine controls the Z-axis guide to move the welding torch to the weld seam position.Software design of image processing is jointly programmed by VS and OpenCV.Part of the data processing uses the data calculation function of MATLAB software.A robot welding positioning system based on bichromatic light source vision is developed for standardized welding workpieces.The image processing algorithm adopts the preliminary welding workpiece.Positioning and precise positioning of the weld seam are implemented in two steps.The dual-color light source effectively improves the discrimination between the recognition target and the background,and the step-by-step image processing algorithm improves the reliability of the positioning algorithm.Secondly,the feature point extraction algorithm based on laser structured light is studied.The triangulation method is used to calculate the corresponding parameters of the welding workpiece.According to the calibration of the camera internal and external parameters,the measured parameters are converted into Cartesian physical coordinate system,through the upper and lower positions.Machine coordination work to complete automated welding.