Research On Visual Guided Welding Starting Point Location Technology For Y-type Pipe Joint

In recent years,with the increasing degree of automation in the welding field,automated welding machines with visual guidance capabilities have gradually become a hot research topic in the intelligent welding field.This article focuses on the demand for the welding starting point guidance function of automated welding machines and conducts research on a machine vision-based intersection line welding starting point target tracking system.The main research contents include: selection of the optimal welding gun pose for intersection line welding,optimization of the algorithm for extracting image feature points for welding starting points,design of an automated welding machine for Y-shaped pipe joints with intersection line welding,and development of a dedicated detection system interface for welding starting points.The main work of this article is as follows:(1)Determine the optimal welding gun pose for intersection line welding.By establishing a spatial rectangular coordinate system and using coordinate transformation methods,the spatial equations of the intersection line weld in both the main pipe and branch pipe coordinate systems are derived.Combining MATLAB for model visualization,the correctness of the intersection line mathematical model is verified.According to the welding process requirements,the welding gun pose model is established through an auxiliary coordinate system,and the optimal welding gun pose for intersection line welding is determined.(2)Propose an improved Canny edge detection algorithm that combines Contrast Limited Adaptive Histogram Equalization(CLAHE)image edge detail enhancement with Otsu’s adaptive threshold segmentation.To address the problem of low edge feature clarity in traditional Canny edge detection algorithms,the improved algorithm can extract clear and continuous edge features of the workpiece by integrating CLAHE image edge detail enhancement with Otsu’s adaptive threshold segmentation.To solve the problem of large pixel coordinate value deviation of the output feature points in the detection process,sub-pixel processing is proposed after edge detection,and the sub-pixel coordinates of the feature points are output for guiding the welding starting point.Average guidance error reduction of 24.97%for sub-pixel guidance results compared to pixel guidance results.(3)Complete the design of an automated welding machine for Y-shaped pipe joints with intersection line welding.Taking the intersection line weld in the MPCVD equipment reaction chamber as an example,the mechanical structure of the automated welding machine is determined by analyzing the path generation method of the intersection line weld,and the mechanical structure design of the machine is carried out using Solid Works.To meet the requirements of the optimal welding process,a welding gun intermittent swinging model is established,and the starting point position and the initial angle of the welding gun are determined.According to the above requirements,the technical indicators of the relevant components are formulated,and a physical prototype is built.(4)Develop a dedicated detection system interface for welding starting points.Based on Py Qt,a dedicated detection interface for welding starting points is developed.To verify the reliability of the system,pixel-level and sub-pixel-level feature point detection are performed at six different spatial positions of the welding gun.The detection results show that the average error of pixel-level detection is 0.4033 mm,with an average error rate of 0.92%;the average error of sub-pixel-level detection is 0.3025 mm,with an average error rate of 0.69%.The guidance error of a single sample is less than 0.5mm,which verifies the reliability of the dedicated detection system.By developing a visual guidance-based welding starting point target tracking system,the feasibility of the proposed welding starting point target tracking method is verified.The experimental results show that the target tracking method has high measurement accuracy and reliability.The research results of this article have important significance for the research on welding starting point target tracking methods for Y-shaped pipe joints.