Research On Vision Sensing And Tracking System For Narrow Gap Welds In MIG Welding

Technical researches have carried out during the narrow gap groove automatic MIG welding process in this paper.Aimed at solving the problems that the weld feature information is not obvious,the image processing time is long,and the system accuracy and stability need to be further improved,the narrow gap weld tracking platform is designed.Based on it,the groove visual image processing and deviation calculation,the system control strategy research,the software and hardware integration and experimental verification are also performed.Break through has been made in the key difficulties,such as the design of narrow gap groove welding seam tracking recognition sensor,the progress of the accuracy and real-time of the joint image processing algorithm under interference conditions such as arc light,splash,smoke and dust.In this project,a tracking system for narrow gap groove welds was developed,a "template matching" image deviation extraction method was designed,and tracking system software was developed.On this basis,real-time tracking tests for the linear and v-groove butt weld line and curve welding of I-type groove butt welds were performed during MIG welding process,the test results meet the accuracy requirements of automatic welding tracking for narrow gap grooves.The laser vision sensor method is used to analyze the characteristics of the groove visual imaging in the narrow gap groove welding process,and the visual sensor device is optimized and improved.By using 650 nm wavelength line laser、improved telecentric lens and matching CMOS high-speed camera,it solves the problems of obstructed vision caused by groove butt welds、unclear visual imaging and affecting image processing,improving the engineering applicability of the system.On the basis of studying the commonly used seam image processing technology,three suitable image processing flow schemes are designed.And the image processing time and image processing effect of each scheme are also compared and analyzed.As a result,the optimal scheme is selected prepared for algorithm optimization.Considering the effects of real-time and accuracy of seam image processing,the previous image processing algorithms such as median filtering、Canny edge detection operator is optimized in this article.To improve and innovate the image processing process for extracting deviations in the later stage,a "template matching" deviation extraction scheme has been designed.By combining multi-threaded parallel processing 、 splitting complex calculation tasks and combining calculation results,the final process running time is controlled within 60 ms,meeting the requirements of real-time control.At the same time,in order to improve the stability of the system,the system fuzzy-PID control model is designed.Finally,high-carbon steel components are used to carry out the real-time weld tracking test of the straight track and the curved track of I-shaped and V-shaped groove butt welds.Tests have shown that for I-type linear joints,when the welding speed is within the range of0.6～0.9m/min,the maximum tracking error is within 0.16 mm for gaps less than 1mm.For Vtype linear joints,when the welding speed is 0.8～0.9m/min,the maximum tracking error with groove gap of 3～5mm is basically within 0.45 mm.And the maximum tracking error is within0.17 mm for I-shaped curved joints,which meets the requirements of narrow gap tracking control accuracy.