Research On Weld Seam Automatic Tracking System Of Spirally Steel Tubes Based On OpenMV

With the full development of industrial 4.0 revolution,welding automation has aroused widespread concern in the whole society.In the process of realizing welding automation,automatic seam tracking has always been a key and difficult problem,and the reason lies in the interference of external environment such as strong light,smog dust,arc during the welding processing.How to accurately identify the position of welding seam through sensors and guide the external equipment has been troubled all the related researchers.This research is based on the requirement of spiral pipe welding automation transformation in Wuhan Iron and Steel Group Co.,Ltd,a visual seam tracking system is designed by using the machine vision module in the market,which is to improve welding productivity and ensure welding efficiency while reducing the production intensity of workers.This paper introduces the development and edge trends of seam tracking research,the composition and production process of submerged arc welding equipment for spiral pipe,and introduces the principle of seam tracking based on the workflow of automatic seam tracking system.It can be divided into three parts: image acquisition,image processing to identify weld information and drive motor to adjust the torch.This research is based on OpenMV module which is programmed by MicroPython.The program is designed to extract seam position information by using the highly expansibility of the module,the direction and distance of torch offset are adjusted by I/O output as motor recognizable instructions.In the experimental process,we mainly verify the seam characteristics of spiral pipe under different illumination intensity and angle conditions,and the influence of different visual acquisition angles on image quality.Through comparison,we can get the best optical conditions.For the image processing part,through consulting the relevant research of scholars,we can determine the core algorithm involved in the processing process.For every step of image processing,we can get through it.By comparing the performance of different methods,such as processing effect and hardware operation efficiency,the final processing scheme is decided.The software and hardware scheme of the system is adjusted and modified by connecting and testing the equipment on the production site.According to the field test results,the light source illumination with high illumination intensity can highlight the straight line characteristics of the weld seam,which is beneficial to weld seam recognition;the median filter can effectively suppress the random interference generated by the environment in the production process,and improve the accuracy of weld seam feature extraction to a certain extent;combined with the results of previous weld recognition,the optimized arithmetic mean filter can effectively prevent the weld seam.The influence of anti-optical noise on tracking system.