Research On Real-time Tracking Method Of Corrugated Plate Fillet Weld Based On Dual-point Laser Vision Sensor

Corrugated sheets are widely used in containers,rail cars,highway bridges,etc.In actual industrial production,most of the corrugated plate fillet welds still use electrode manual arc welding and mechanical profiling semi-automatic welding.Today’s more advanced method is that welding robot using laser stripe vision sensor to scan the weld before performing automatic welding.In addition,there is a laser vision sensor scanning in the front,and the welding robot lags the sensor a ripple to weld.The above two methods will affect the welding quality due to the welding deformation generated during the welding process.There is also advanced detection method for laser stripe vision sensing.The vision sensor is fixed on the welding gun,and the laser stripe irradiation point,that is,the detection point of the vision sensor,has a certain forwardlooking distance from the welding point.In this detection method,the front sight distance of the laser vision sensor will change at the bending of the corrugated plate,and the weld seam deviation caused by the change of the front sight distance is not the weld seam deviation corresponding to the real position.In addition,since the posture of the welding torch changes at the bend of the corrugated plate fillet weld,the weld moves out of the sensor’s field of view,and the tracking and detection of the entire fillet weld cannot be realized.In actual welding production,factors such as welding deformation,manufacturing errors of corrugated plates,and environmental influences will continuously change the distance from the welding torch to the position to be welded.In order to the welding torch always be aligned with the position to be welded,it is necessary to detect the deviation of the weld seam in real time.According to the change of the welding seam deviation,corresponding adjustments are made by controlling the welding torch to ensure stable and reliable welding quality.Therefore,there is an urgent need for a real-time welding seam deviation detection and automatic tracking method to realize the welding production automation of corrugated plate fillet welds,guarantee welding quality improve production efficiency,and meet the needs of various industries.This paper takes the trapezoidal corrugated plate fillet weld as the research object,and builds an automatic welding seam tracking system based on two-point laser vision sensing.The hardware and software sensor technology for dual-point laser position sensing are developed.In order to realize the real-time detection of corrugated plate fillet weld welding as the research goal.The main components and working principle of the welding seam automatic tracking system are introduced.The system software is designed and developed based on Visual C++ visual programming language.The software has comprehensive functions,simple interface and simple operation.It has image acquisition to memory,memory image data copy and backup,memory image data filtering,image data binarization processing,laser spot image shape analysis,rationality judgment,laser spot binarization,curve drawing,welding seam deviation data calculation,servo drive signal processing,RS485 communication data processing,control robot start welding,welding stop and other functions,provide a reliable data source for subsequent data analysis.A two-point laser vision sensor welding seam tracking detection method is proposed.A dual-point laser vision sensor is designed.The two-point laser light sources of the sensor are installed on both sides of the welding torch.The axis of the laser light source is parallel and coplanar with the axis of the welding torch.The two laser spots A and B coincide with the projection of the end of the welding wire in the fillet weld distribution direction of the corrugated plate.This feature of the sensor is that the welding point is synchronized with the detection point,which realizes the real-time detection of the automatic tracking of the welding position of the welding point.Through the principle of geometric optical imaging,the mathematical expression between the lateral and longitudinal deviation of the weld and the displacement of the image on the CCD photosensitive element is deduced,and the practicability of the sensor designed in this paper is verified theoretically.The checkerboard is used to calibrate the actual size of the unit pixel of the vision sensor.It is calculated that the use distance of the laser vision sensor is 150 mm,and the actual size corresponding to the horizontal and vertical unit pixels within its field of view is not greater than 0.1mm,which has higher resolution.The acquisition,data processing and mathematical analysis model of welding torch position sensing signal(laser spots)based on double-point laser vision sensing are established.The characteristics of the laser spots are analyzed when the welding seam has horizontal and vertical offset,lateral offset,vertical offset,flat plate direction offset,and corrugated plate direction offset.The direction along the welding torch is the longitudinal direction,and the direction perpendicular to the welding torch and the welding seam is the transverse direction.Through the geometric position relationship,the mathematical relationship between the offset Ad of laser spot A on the corrugated plate and the offset Bd of laser spot B on the plate,the transverse offset Hd and longitudinal offset Vd of the weld,and the angle β between the welding torch axis and the corrugated plate are derived.The mathematical model of eigenvalue analysis is established.In this way,the horizontal and vertical deviation values of the welding seam can be calculated,and then the cross electric translation stage installed between the welding gun and the sixth axis of the welding robot can be controlled to move the welding gun,for realizing the real-time deviation correction of the welding seam.The numerical relationship between the long axis of the ellipse of the laser spots on the corrugated plate and the flat plate detected by the vision sensor and the angle βbetween the welding torch and the corrugated plate is analyzed,which can further realize the detection and control of the angle between the welding torch and the corrugated plate or the flat plate.The laser spot analysis of the bending section of the fillet weld of the corrugated plate is carried out,and it is found that in the straight section and the bending section of the corrugated plate,the short axis of the laser spot on the corrugated plate projects Vsd in the direction of the vision sensor is a constant value,and the long axis of the laser spot on the corrugated plate is only affected by the angle between the axis of the laser light source and the corrugated plate.Therefore,when calculating the deviation of the weld seam,the entire or half arc contour of the laser spot can be used to compare with the reference position.Therefore,when processing the image,focus on processing the entire or half arc contour of the laser spot.For the influence of welding arc and spatter,narrow-band and band-pass filtering methods are used to reduce the influence of arc light,software technology is used to eliminate the influence of welding spatter on the image,and median filtering is used to reduce the image noise signal.The image binarization method of the clustering algorithm is used to process the weld image to obtain the edge of the laser spot.About the characterization of laser spot position,according to the characteristics of the position change of the laser spot when the welding seam deviation occurs,it is determined that the sum of the Yaxis data of each point on the semicircular edge of the laser spot is used to obtain the average value,and the average value is used as the characteristic value to detect the welding seam deviation.About automatic tracking test of corrugated plate fillet weld,the laser spot image position and laser spot image data are analyzed when the welding torch moves horizontally and vertically.When the welding torch moves,the position of the laser spot image and the binarization curve of the edge of the upper semicircle change.Through the change of the position of the laser spot,the direction change of the welding seam deviation can be judged.The test collects the data of each point on the semicircle edge of the laser spot when the welding torch moves horizontally and vertically from﹣4.5mm～﹢4.5mm,and draws its binarization curve.The eigenvalues of the Y-axis of the semicircle edge on the laser spot are calculated when the welding torch moves0.5mm in each step,eigenvalue error curves are plotted and the sensor detection error is less than 0.2mm,and the welding seam tracking detection accuracy of the dual-point laser vision sensor is 0.49 mm.The trapezoidal Q235 galvanized corrugated plate is selected,the thickness of the plate is 1.5mm,the wave height is 80 mm,and the angle of the corrugated plate is 120°.Under the appropriate welding process parameters,using the dual-point laser vision sensing automatic welding seam tracking system which is built in this paper and the corresponding detection method,the welding tracking test is carried out.The welding seams with beautiful surface and good forming quality are obtained.The automatic tracking and detection of the weld deviation of the corrugated plate fillet weld is well realized.