Research On Automatic Location Of Whitebody Solder Joint Based On Machine Vision

In the automotive manufacturing industry,resistance spot welding,as the most widely used welding technology for body-in-white,is of great significance for maintaining the rigidity and quality of the entire vehicle.Many automotive OEMs have already achieved robotic automated spot welding production.However,the welding spot quality inspection method corresponding to this method is not very automatic.Currently,most of the solder joint inspection equipment currently used can only use artificial welding points.Performing off-line sampling inspections has low detection efficiency and can not timely feedback the quality of solder joints and cannot adapt to the trend of industrial production automation.In this paper,based on the existing ultrasonic solder joint quality detection technology,a method for automatic identification and location of white body solder joints based on machine vision is studied to quickly locate the solder joints and guide the probes of the robot end solder joint inspection equipment to reach the solder joints for quality.Inspection to realize the automatic quality inspection function of solder joints.The main research content of this paper is as follows:(1)Build a machine vision test platform based on the needs of the solder joint location function.Using a certain type of vehicle body in white as the test object,hardware selection of machine vision systems,including cameras,lenses,light sources,robots,industrial computers,etc.,and reasonable arrangement of the hardware;designing welding spot identification and positioning process,according to the selection of The hardware selects the driver interface,the appropriate programming language,and the function library to determine the program architecture.(2)According to the characteristics of solder joint images,the image processing positioning algorithm for solder joints is designed.The characteristics of solder joint image contours and gray-scale distribution were analyzed.Image processing algorithms including preprocessing methods such as image filtering and edge detection were designed.Hough transform principle was used to fit the solder joint profile to solder joint location,and the solder joint was designed according to the gradient curvature characteristics of solder joints.A multi-contour fitting andclustering analysis method has a good effect on the positioning of the center of irregular solder joints.(3)Solve the camera model parameters,study the robot guidance method,so that the ultrasonic detection probe at the end of the robot is aligned with the center of the spot welding spot.Analyze the principle of the camera model,calculate the camera's internal parameter matrix,distortion coefficient,etc.,and use the principle of affine transformation to calculate the conversion method from the camera coordinate system to the robot's world coordinate system.The robot interface communication method is studied to realize the real-time information exchange between the industrial computer and the robot,so that the industrial computer can guide the robot to move to the position of the solder joint.(4)Based on the functional requirements of body-white solder joint positioning and quality inspection,a set of software has been developed to realize functions such as image processing,image data storage and setup,guided robot movement,solder joint quality detection,and the like;Testing,repetitive positioning inspection tests on solder joints in different locations,analysis and summary of test results and related data,verification of the effectiveness of solder joint identification and positioning methods,and the implementation of automated inspection of white body solder joints on the production line Important reference value.