Research On L-Shaped Plate Weld Identification And Tracking Technology

In recent years,with the increasing market share of China’s automotive manufacturing industry globally,the requirements for welding technology in the automotive manufacturing industry have also increased.Among them,L-shaped plate is an important component in automotive equipment manufacturing,widely used in automobiles as supporting parts.Traditional manual welding has many drawbacks,such as positioning deviation,which can easily lower product quality and fail to meet production requirements.Manual teaching welding also requires certain experience and technical skills in robot control.Therefore,laser vision-guided robot welding,as an advanced technology,is gradually replacing manual teaching operations to improve production quality.By using vision sensing technology and welding tracking technology in combination,intelligent and visualized robot welding can be achieved,leading to improved welding efficiency and quality.Therefore,the research on L-shaped plate welding seam recognition and tracking technology is of significant value,with the following specific contents:Firstly,the framework of the welding system for L-shaped plate is introduced.The design principle of the laser vision sensor is detailed,and the component selection and construction of the vision platform are completed according to the requirements of the vision system.The robot gripping system and the model and functions of the robot welding system in the L-shaped plate welding framework are briefly introduced.Then,based on the imaging principle of the camera,camera intrinsic and extrinsic parameters are analyzed and modeled.On this basis,by combining line structured light with industrial camera using the least squares method,the structured light plane is fitted,and the hand-eye relationship matrix is solved through rotation and translation,completing the hand-eye calibration of the robot system and determining the transformation relationship between spatial coordinate systems.Through experiments,it is proven that the proposed calibration scheme can accurately identify the feature point positions of the L-shaped plate weld seam.Subsequently,in order to accurately extract the feature points of the weld seam and overcome the problem of reduced precision positioning of the L-shaped plate weld seam by visual system due to external disturbances and noise,the original laser weld seam image collected is optimized using software platform.ROI image processing is performed first to extract the main feature region of the weld seam image,and then the weld seam image is filtered to optimize the noise interference in the image using the maximum inter-class variance method.The laser lines are segmented from the background plate of the image.Morphological processing is then applied to eliminate the inner holes in the laser lines,resulting in a continuous and clear laser stripe image.The obtained stripe image has a certain width and cannot be directly used for feature point calculation.Therefore,the center line of the laser stripe is extracted using the grayscale centroid method to obtain the centerline.Finally,the feature points are extracted and fitted into the required welding path using the least squares method.Finally,based on the tracking system of the L-shaped plate weld seam experimental platform,the path planning of the L-shaped plate is completed.The specific process of weld seam tracking experiment is determined,and the offline tracking method is used to achieve trajectory tracking of the teaching path.Experimental validation is carried out,and the results show that the designed L-shaped plate welding system in this paper can replace the teaching pendant for guided welding and complete the weld seam recognition and trajectory tracking of the L-shaped plate.