Research On Weld Recognition And Automatic Tracking System Based On Vision Sensor

Thin plate structure is widely used in many manufacturing industries such as shipbuilding,mechanical processing and aerospace,and thin plate splicing adopts welding technology.Due to the increasingly serious problems of unstable welding quality,low productivity and varying levels of welder skill in manual welding operations make automated welding the main development direction of modern welding manufacturing technology.The essential factor of automatic welding is the welding seam identification and weld tracking.The vision sense technology boasts the advantages of strong repeatability,rich information,without contacting with the workpiece and high measurement accuracy,being applied in research on the welding seam identification and tracking by scholars from domestic and foreign.Affected by many different kinds of factor in the automatic welding process,the welding gun possibly deviate from the designed weld path.It is necessary to use a suitable welding seam tracking algorithm to correct the movement path of the welding torch.Therefore,the paper conducts a study related to the image processing algorithm of welding seam recognition and the algorithm of weld tracking.This paper first designs a general scheme of a vision sensing based weld seam identification and automatic tracking system.The author selects the key equipment of the system according to the actual needs of the project,and then designs the internal structure of the laser vision sensor.Then,the author as well as calibrated the camera in the laser vision sensor to establish the mapping correspondence between the image coordinate system and the world coordinate system.By analyzing the grayscale distribution characteristics of the acquired images.In this paper,a gray scale stretching algorithm is used to enhance the contrast of the acquired image to propose a Scharr operator-based edge detection algorithm for acquiring ROI regions.An adaptive directional template method is proposed in the paper for extracting the centerline of the ROI region.The method uses the second-order difference method and the polar search method to extract the weld seam feature points so as to achieve weld seam recognition.The experimental results of weld seam recognition are analyzed to prove the feasibility of the used image processing algorithm.The weld seam position data of a section of the weld is interpolated using the cubic spline method to ensure the smooth movement of the welding gun during the welding process.For the thermal deformation problem during thin plate welding,this paper proposes a combination of PID control and double-queue control algorithm for automatic weld seam tracking based on the stepwise retreat weld method for planning the torch motion path.By analyzing the results of two sets of weld seam tracking tests,the paper concludes by demonstrating the reliability and accuracy of the proposed weld seam tracking algorithm.