Space Weld Seam Recognition And Tracking Based On Line Structured Light Vision Sensor

With the development of industrial technology towards automation and intelligence,the requirements for welding automation in industrial manufacturing are gett ing higher and higher.The traditional welding robot adopts the method of teaching and reproducing to realize automatic welding.However,due to the variable working position and complex working environment in shipbuilding,it is difficult to apply the met hod of teaching representation in shipbuilding.Therefore,aiming at the problem that it is difficult to adopt automatic welding in shipbuilding,this paper takes the watertight patching plate welding in shipbuilding as the goal,and carries out the research on the the 3D weld tracking system in the workpiece robot calibration environment.Based on the requirement of weld tracking system,a laser vision sensor which can obtain the weld position information in real time is studied in this paper.Combined with the imaging principle of the camera,the accurate measurement model of the laser vision sensor mapping from 2D coordinates to 3D coordinates is derived,and the transformation from pixel coordinates to 3D coordinates is realized.The calibration algorithm of the laser vision sensor is studied,combined with Opencv to achieve fast calibration.Finally,the hand-eye calibration of the robot is studied,and the transformation from the two-dimensional coordinates of the image to the three-dimensional coordinates of the robot base is realized.Some researches were carried out to solve the problem of difficulty in extraction of weld features caused by the interference of splashing,arc light,and pre-solder joint during welding process.Using the continuity of weld features in time and space,weld feature tracking is used instead of extraction of weld features.In the welding process,the welding seam features were extracted by morphological method only in the first frame when the welding power supply was not turned on.In the subsequent frames,two parameters were used to update the tracker with different rules to track the characteristic area of the welding seam,and the two sets of data were fused by the Kalman filter.The robustness of the welding seam tracking system is improved,and the cumulative error of the tracking system is reduced,and the welding precision is improved.To avoid curve welding,the welding pose can not be adjusted,resulting in welding failure.The on-line calculation method of welding pose is studied.By establishing the coordinate system on the welding path points,the robot base coordinate system and the coordinate system established on the welding path points overlap,so that the welding attitude can be adjusted in real-time during the welding process.Improve the adaptability of the weld tracking system.Finally,according to the actual welding environment of the watertight patching plate in shipbuilding,the overall scheme of the weld seam tracking system is designed.The welding system is divided into three parts: the starting point and end point acquisition module,the image acquisition and processing module,and the task execution module.The hardware selection and software design are also carried out.Finally,relevant experiments are designed according to the method proposed in this paper.The experimental results verify the feasibility of the method in shipbuilding.