Research On 3D Measurement Of Welding Seam Between Towers And Flanges’ Ribs Based On Light Section Method

The power tower is an important part of the power transmission line.The steel tube tower is widely used because of its structure has great appearance,small figure coefficient and easy to install.At present,the neck flange connection is usually used in the multi-section connection of steel tube tower in China.During the production and manufacture,the flange and its rib plate should be welded to the towers’ body.The size and specification of steel tube tower are various,the tower body is usually ccylindrical and polygonal,and there are many ribs along the flange circle.The location roller which used to supporting the tower have large errors duting the rotating,so manual welding is still used,which has the problems of high labor intensity and low efficiency.The real-time location of welding seam is the first problem to realize the flange rib plate of tower’s automatic welding.In this paper,a method of seam 3D feature recognition based on the light cut plane scanning and 3D point cloud processing is proposed in this paper.The main work includes:(1)According to the welding process and accuracy requirements of the flange rib plate of power tower,combined with the three-dimensional measurement characteristics of line structural light,a three-dimensional measurement model of the flange rib plate of the power tower is established by the Light section method method.The three-dimensional coordinate relationship between the image point and the corresponding point in the plane of the structural light under the camera coordinate system is obtained,and the 3D measurement of the object is realized.(2)The parameter calibration method of the three-dimensional measurement system which based on optical cutting method is studied.The camera is calibrated by Zhang’s calibration method.The integrated calibration method is designed based on the principle of constant cross ratio to realize the solution and calibration of the structural light plane.The feature point coordinates are calculated according to the plane equation,and verify the fitting accuracy of the structured optical plane,and the calibration software of the measurement system is developed to realize the calibration of the system’s parameter.(3)The characteristics of laser fringe image are compared and analyzed.Design the image processing flow and algorithm to extract the structure image’s center.The efficiency of the extraction is considered when the effective information of laser fringe is kept to the maximum extent.The paper uses Zhang’s skeleton refinement algorithm to extract the structural light center,calculates the direction information of each point according to the neighborhood points,and realizes the sub-pixel extraction of the fringe center by combining the gray gravity center method.(4)According to the geometrical characteristics of weld space,a 3D feature extraction algorithm based on point cloud is designed.Firstly,the point cloud is de-noising by radius filtering,and the point cloud is simplified by voxel grid method.The plane segmentation is realized by using the uniform plane segmentation method based on RANSAC,Used the edge feature extraction algorithm which based on neighborhood distribution extract the edge.Finally,used the multi-step feature extraction method to extract the weld feature point cloud.Build the hardware platform and developed the software system to verify the effectiveness and feasibility of the algorithm.The measurement results are compared with the real data,and the dimension error of the 3D extraction results is less than 5%,which satisfied the error requirements of welding process.