Research On Welding Robot System Of Forming Shoulder Based On Stereo Vision

Forming shoulder is a key component of bag forming on vertical packaging machine,and its manufacturing accuracy directly affects the packaging speed,packaging form and quality of vertical packaging machine.The current welding of the forming shoulder is mainly done manually,and the welding quality is highly dependent on the experience of the welders,which makes it difficult to guarantee the consistency of the product,so the development of a forming shoulder welding robot system is a practical need for the relevant enterprises to achieve automation upgrade.In this paper,a welding robot system is designed independently for the automated welding of forming shoulder,and the main research contents and conclusions are as follows:(1)According to the welding requirements of the forming shoulder,a set of forming shoulder welding robot system was designed.First of all,the analysis of the welding requirements of the forming shoulder,and according to the welding requirements of the overall design of the forming shoulder welding robot system,including the independent design of a set of fixtures to achieve rapid and high-precision clamping and positioning of the forming shoulder;the use of six degrees of freedom collaborative robot arm plus a single degree of freedom of the displacement machine welding The robot configuration of six-degree-of-freedom collaborative robot arm with a single-degree-of-freedom transposer is used to ensure the accessibility of the torch position;the stereo vision measurement scheme of line laser monocular vision is determined to realize the actual surface measurement of the parts to be welded and the weld seam measurement scheme.Finally,the hardware design and software design of the welding robot system were completed according to the overall scheme.The hardware part completed the selection,inter-module communication design and the system argon arc welding anti-interference design;the software part determined the overall architecture and completed the control software design.(2)The line laser monocular stereo vision system was studied,and the 3D reconstruction of the surface of the weld to be made and the pre-identification and 3D measurement of the weld seam were realized by stereo vision.First,the model of the line laser monocular stereo vision was established,and the vision system was calibrated.Then,the algorithms related to the line laser image processing were studied,and the accurate extraction of the laser stripe center was achieved by the proposed improved multi-threshold laser stripe region segmentation algorithm,multi-scale adaptive laser stripe center extraction algorithm,and reflection interference removal algorithm based on maximum deterministic extrapolation.Then,outlier rejection and downsampling of the point cloud are achieved by statistical filters and voxel raster filtering.Finally,the pre-recognition and 3D measurement of the weld seam is achieved by using the corner point feature of the point to be welded,combined with the area prediction of the point to be welded by Kalman filtering.(3)The automatic programming system of the forming shoulder welding robot was studied.First,the kinematic model of the six-axis robot arm and the displacer was established based on the D-H parameter method,and the kinematic models of each device were correlated to establish the kinematic model of complex spatial curves welding.Then,to solve the problem of difficult to accurately capture the trajectory of complex spatial curves,NURBS curves are introduced to smoothly approximate the trajectory described by the 3D point cloud,which can obtain a smoother and closer curve to the original trajectory compared with the traditional interpolation method.Finally,the trajectory planning of the welding process is studied and the interpolation of the spatial surface welding attitude based on quadratic spherical linear interpolation is implemented.(4)The forming shoulder welding robot system was built and experimentally verified.First,the forming shoulder welding robot system hardware platform and software system was built to read into the 3D point cloud file to automatically generate the welding job file,manual control of the robot and other equipment and 3D visual motion simulation and other functions.Then,the line laser monocular stereo vision system and automatic programming system were tested experimentally.The test results show that the line laser monocular stereo vision system for the Z-axis measurement error of the standard measuring block is less than 0.088 mm,and the automatic programming system functions properly,in line with the design expectations.Finally,the process parameters of welding were explored through process tests,and automatic welding experiments were conducted.The experimental results show that the system can realize the welding actions of the forming shoulder,in line with the design expectations.