Design Of Automatic Welding System Based On Visual Guidance

Welding technology as a kind of important in industry processing technology,has important effects on every industry Welding technology development so far,the main can be divided into manual welding semi-automatic welding and automatic welding of three types In the actual welding process of a large number of soot and high strength arc,great harm to human body health With the advent of the computer era,in order to reduce the harm to human body,welding technology is also developing in the direction of comprehensive non-contact measurement.Non-contact welding method can greatly improve the working environment of workers and reduce the interference of human factors on welding accuracy In recent years,with the improvement of mechanization and the gradual development of welding robot technology,automatic welding robot based on vision guidance has gradually become a research hotspot.Therefore,this paper adopts the principle of laser triangulation.Firstly,according to the requirements of the system,the hardware composition and the control relationship between each part of the system are determined.Then,the hardware of each part is introduced successively.Then,the hardware control system and its communication scheme of the welding system are described,and the hardware platform of the whole system is built successfully.In system calibration process first for linear structured light measurement principle and camera imaging principle has carried on the theoretical derivation,then the camera calibration and line structured light plane calibration experiment,by zhang method to calculate the calibration parameters,finally expounds the manipulator coordinate system and the relationship between the camera coordinate system,prepare data for weld center.Secondly,in view of the problem that the automatic welding system determines the initial position of the micro-weld of spliced steel plate,the region with the highest similarity to the template is the initial area of the weld through the template matching calculation,and then its coordinate position is determined through the matching algorithm.Then,the improved STORB algorithm is used to detect the feature points in the initial area of the weld and match them with the template one by one.Then,the matching points are optimized by RANSAC algorithm to eliminate the mismatched point pairs.Finally,the precise positioning of the initial point of the weld is realized according to the initial point determination criterion and error analysis.Third,according to the welding process,spraying,arc,soot and other factors and outside conditions such as diffuse and specular reflection interference,reduce the detection area is extracted using ROI area,improve recognition rate,and then adopts frequency details enhancement approach improve structured light stripe and the overall comparison,then carries on the segmentation of laser stripe,and by using morphology processing to remove scatter noise,Then the center line of the light stripe is extracted.Finally,STORB algorithm is used to detect the breakpoint of the weld zone of the center line and obtain the coordinates of the final center point of the weld.Finally,the supporting software of the automatic welding system was built on the Pycharm programming platform,and the three main interfaces were introduced successively.Finally,the actual measurement tests of different welding parameters were carried out on the stainless steel plate stitching welds through the automatic welding system built,and the feasibility and effectiveness of the whole system were verified.