Weld Identification And Trajectory Planning Based On Machine Vision

With the "Made in China 2025" programmatic is documented,China gradually into the manufacturing power.China’s industrial manufacturing industry is developing rapidly,and the requirements of welding also will be higher and higher.In the traditional manual welding process,there is a low welding efficiency and welding quality is not guaranteed and other production problems.At the time poor welding environment and labor intensity issues seriously affect the health of workers.The current welding robots have been widely used in industrial applications.However,most of them are based on the teaching mode of welding robot,This type has the advantages of high stability,simple operation and repeatable.It can replace manual welding to a certain extent.But there are very serious efficiency problems.In the process of teaching,It need to teach the number of points more often this is very time-consuming,poor flexibility and does not have strong robustness.These problems are the technical difficulties of welding robot application technology,which seriously restricts the popularization and application of welding robot technology.With the development of image processing and computer vision technology,robots with visual sensing feedback system are gradually used in the field of welding.The visual feedback system can effectively avoid the various shortcomings of robot teaching,and track and measure the weld in real time.And,in the measurement process,the weld deformation and irregularities can be captured in time,with high flexibility.In this paper,the trajectory planning based on computer vision and the trajectory planning of welding robot are taken as the research object.By improving the image processing of the weld,the ability of the computer to identify the weld and improve the operation efficiency.At the same time,to improve the trajectory planning of welding robots,so as to achieve the optimal operation based on time,so as to further improve the operating efficiency of the entire welding arm.This paper studies image preprocessing techniques.For the traditional image processing algorithm in dealing with the interference with the weld image when the center line image intermittent or deviate from the center line and other issues.In this paper,we propose a new algorithm based on improved pre-processing of dark channel,which combines filter denoising,threshold processing,morphological operation and edge detection to make the algorithm robust and clear,Centered and continuous weld centerline.Which can obtain the characteristic point coordinates of the centerline of the weld image by the detection of the feature points to guide the welding robot to finish the welding.In this paper,the feature matching technique of two weld images based on binocular vision detection is studied.Brute-Force matching algorithm based on SIFT feature extraction and Flann matching algorithm based on SURF feature extraction.Depth analysis of the structure of the algorithm,clever design experiments to verify the feasibility of the algorithm to get a time and match the correct rate can meet the needs of welding robot feature matching algorithm.In this paper,we study the optimal trajectory planning problem of manipulator.Trajectory planning of a manipulator for a 3-5-3 polynomial interpolation robot arm based on a genetic algorithm.While taking into account the stability of the operation,reduce the swing to ensure that the welding robot in the welding process safe,stable and efficient operation.This paper designs a gesture recognition system based on machine learning,which can quickly and effectively identify the gestures of front-line operation technicians,so as to ensure the safe and effective processing of the sudden occurrence in the production process.This paper designs the man-machine interaction interface of welding robot.it can allow supervisors to more intuitive observation of the robot welding situation,and can manually adjust the image processing process in some of the parameters.So that the welding robot system has a harmonious human-computer interaction environment,and further improve the welding production efficiency.