Research On Multi-Layer Multi-Pass Intelligent Row Welding System Based On Laser Visual Sensing

Medium and thick plate components are widely used in shipbuilding,aerospace,automobile manufacturing and other fields,compared with thin plates,the weld groove of medium and thick plate components is larger,the welding time required is longer,and the welding heat is more serious.If only one layer is welded,the welding effect will be far from meeting the requirements.If the use of multi-layer multi-pass welding method to weld this kind of medium and thick plate workpiece,can greatly improve the effect of welding,at present in the actual welding application,multi-layer multi-pass welding method is basically completed by workers to complete the welding work,which also makes the welding work intensity increase,the welding environment is more harsh,welding quality stability can not be guaranteed,so it is urgent to develop a robot to replace manual welding operations of automated stable welding methods,It is used to solve the problems faced by manual welding above.At present,the welding robot is mainly taught to reproduce the working method,this working mode can only complete the welding task under specific conditions,for the complex and changeable welding conditions,this welding method can not complete the welding task well In view of the problems raised by the appeal and the actual welding needs,this paper designs and develops a laser vision sensing system that can work well with the welding robot,including laser sensor,image grabber,industrial control machine,this system helps to improve the automation level of welding robot in the components of medium and thick plates.In view of the problems raised by the appeal and the actual welding needs at present,this paper will design and independently develop a set of laser structured light vision sensing system that can be well adapted to welding robots,which is mainly composed of the following parts: laser sensor,image grabber,industrial control machine.The completion of this system will improve the automation level of welding robots in medium and thick plate workpieces to a certain extent.In this paper,the feature information is extracted from the weld image containing the groove feature information of medium and thick plates collected from the laser vision sensing system,and it is used to calculate the bevel weld size information later.In this paper,a set of image processing algorithms for groove weld images of medium and thick plates is designed,which filters the image of the weld to filter out the noise influence on the original image,and then performs image binarization on the weld image processed by the filtering operation to make the weld image highlight from the background,and then extracts the weld centerline of the doubled image to make the weld feature information more obvious,and finally performs the least squares line fitting and merging of the extracted centerline to calculate the intersection of two or two straight lines between the three lines.The resulting intersection point is the feature point we require.For the three-dimensional coordinate restoration of weld feature points,a camera parameter calibration method based on line structured light is proposed.It can realize the restoration of weld feature points from two-dimensional pixel coordinates to threedimensional world coordinates.Experimental verification shows that the error in the two directions of the X axis and the Y axis is less than 0.7mm,and the error in the Z axis direction is less than 0.5mm,which meets the actual welding requirements.