Research On Laser Displacement 3D Reconstruction Measurement Method For Weld Surface Defects

Welding defects will be produced due to welding conditions and environment.Welding defects are distributed inside or outside the welds,and the welding surface defects account for a large proportion,such as dents,undercuts,cracks,incomplete welding and so on.To ensure welding quality,weld inspection is necessary.The line laser scanning detection technology for weld surface defects has become mature,which has high detection efficiency and can determine the position of weld surface defect,and the defect is that the geometric measurement precision of weld surface defect is limited.The laser scanning measurement method is used for post welding detection.The size and shape of the welding defects are obtained by measuring the height difference of the weld surface,and the quality of welding can be evaluated to meet the industrial requirements.At present,the application of line structured light to weld surface defect detection is widely applied,and there are many mature products,but the detection accuracy needs to be improved.In view of the limitations of the existing line laser scanning detection,a detection method of weld surface defects based on point laser scanning is proposed to improve the detection accuracy of weld surface defects.The point laser displacement sensor used in this paper has high measuring accuracy,and the three-dimensional information of weld surface defects is obtained by scanning the weldments to be tested with the moving platform.The triangulation method is the optical path geometry principle of measuring the height of the object surface by the point laser displacement sensor,which is divided into the direct laser triangle method and the oblique laser triangle method.The two measuring principles are explained in this paper.The vision detection based on point laser is to reflect the laser point on the surface of the object by the sensor to the photosensitive element,and the surface height is converted from the photoelectric signal.The change of the surface height of the object to be measured is obtained with the moving of the object to be measured.Because of the noise interference in the original data points of the weld surface of the point laser sensor,the noise reduction of the defect data points of the weld surface is processed by the filtering method.Polynomial fitting and BP neural network fitting are used to fit the two-dimensional contour of the weld surface defect,and the advantages and disadvantages of the two are compared,and the two dimensional contour curve and the fitting data are obtained.In this paper,the three-dimensional reconstruction of the weld surface is carried out by the idea of triangulation,which extends from two-dimensional triangulation to three-dimensional triangulation,and extends the plane triangular mesh to the triangular mesh of the space surface.The point of space coordinate data of weld surface defect is obtained by the point laser weld surface defect detection platform.The 3D model of weld surface defect can be reconstructed by Delaunay triangulation algorithm and BP neural network fitting,and the size of the weld surface defect is detected quantitatively.The visual inspection of the test fruit table Ben Q on the point laser sensor is detected.Measurement is a high accuracy method for weld surface defect detection.