Objective:In order to solve the problem that the cross-section profile obtained by the linear structured light vision sensor was not perpendicular to the main direction of the weld and couldnot fully reflect the true morphology of the weld cross-section,a cross-section extraction method of the weld based on the STL model was proposed.Methods:Based on the original discrete point data collected by the line structured light profile sensor,the contour segmentation was carried out by the RANSAC algorithm,the feature points and the edge of the weld were extracted,and the main direction of the weld was determined.The three-dimensional point cloud of the weld surface was spliced by using the cross-sectional contour point of the weld after segmentation.On this basis,the Poisson algorithm was used to fit the surface of the point cloud to construct the STL model.Finally,the STL model was layered based on the position information of the triangular patch to obtain the cross-sectional profile of the weld.Results:The proposed method could accurately extract the cross-sectional profile of the weld by processing the point cloud data with or without defects.Compared with the traditional method,the average cross-sectional profile deviation of the defective weld obtained was reduced by 46％,the maximum value was reduced by 28％,and the standard deviation value was reduced by 41％.Conclusion:The proposed method could effectively extract the cross-section of the weld,which could solve the problem that the cross-section profile obtained by the linear structured light vision sensor was not perpendicular to the main direction of the weld and couldnot fully reflect the true morphology of the weld cross-section,and could provide important data support for subsequent weld defect identification,damage assessment and repair.
维普
万方数据