Study On Arc Shape And Droplet Transfer Behavior Of Rotary Arc Sensing GMAW Flat-welding Based On Image Processing

With the rapid development of the welding industry,welding automation and intelligence are important trends in the future development of welding technology.As an efficient welding method,rotating arc sensing GMAW has gradually become a hot topic and focus of scholars at home and abroad.In this paper,a rotating arc sensing GMAW welding image acquisition system based on backlight is established.The image processing technology is used to study the arc shape and droplet transfer behavior of the rotating arc sensing GMAW flat-welding.In order to analyze the arc shape of the rotating arc sensing GMAW flat-welding,firstly,the arc region is selected by using the ROI extraction,and the image enhancement algorithm for image multiplication is proposed for the weak edge of the arc.Then the Otsu method is used to automatically segment the arc image by segmentation.The pixel area with the value of one is superimposed to obtain the arc area,which improves the arc edge extraction accuracy,and extracts the more accurate arc edge and area.The arc length,arc area and current change of the rotating arc are analyzed,the arc characteristics at different positions of the rotating arc are obtained.For the analysis of the wire extension and arc length variation of the welding wire during welding,an image processing algorithm for the wire extension and arc length extraction of the rotating wire is proposed.The front view of the welding wire at different positions is obtained by perspective transformation,and the welding wire is characterized to obtain its horizontal and vertical coordinates.The image is then cropped,offset,and superimposed,and the wire extension information is obtained by enhancing the grayscale difference between the wire feature and the background.The arc area is selected according to the arc feature localization,and the arc length information is obtained by performing y-direction gray distribution statistics on the arc binary image obtained by the image pre-processing.Finally,the corresponding wire extension and arc length models are established respectively,and the correlation between wire extension and arc length change is analyzed.In order to study the transition behavior of droplets,image filtering was firstlyperformed on the collected droplet images.For the problem that the collected droplets were dark and the boundary was not clear,a combination enhancement method was proposed.Firstly,the overall brightness of the image is enhanced by wavelet transform,then the contrast of the image is enhanced by multiplication method.Finally,the edge of the droplet is extracted by the canny operator,and the least squares fitting is performed on the partially missing droplet edge to obtain the equivalent radius of the droplet,combined with the force of the droplet to analyze the transition shape and transition path of the droplet,the droplet transition frequency is obtained by calculating the time interval of the droplet transition image.