Robotic Visual Positioning System To Detect The Quality Of The Solder Joint On Vehicle Body

Welding is an important link in vehicle production. Thousands of solder joint on a vehicle body play a crucial role in the stiffness, safety performance, load capacity and traveling comfort. But due to several factors such as welding electrode wear,sheet metal surface with dirt and so on, the quality of solder joints on vehicle body are always affected. In order to guarantee the performance of the motor vehicle, the quality of solder joints must be tested in the process of vehicle production. However,the current quality test methods of solder joints on vehicle body whose automation and intelligent level is lower, have been unable to meet the needs of vehicle production, so there is an urgent need for this situation to be changed.In this thesis, robotic technology was introduced into the vehicle-body solder joints quality detection field, combined with advanced ultrasonic solder joint quality detection methods, and the detection accuracy can be guaranteed, besides, the automation and intelligent level of vehicle-body solder joints quality detection can be greatly improved. In the process of vehicle-body solder joints quality detection with robotics, positioning solder joints is the key step. So, visual positioning system of the vehicle-body solder joint quality detection robot is the research focus in this thesis.The main contents are as follows.⑴A robot system platform was built. In this system, MOTOMAN- MH12 robot from Japan yaskawa company was chosen as robot ontology, and the communication between the system was completed with Ethernet. Besides, software architecture of the robot system was built.⑵ According to the outer contour of a vehicle-body solder joint approximates a circle, acquiring the coordinates of the vehicle-body solder joint outer contour center in the world coordinate system was the task in the visual positioning system of vehicle-body solder joint quality detection robot, and some relevant design requirements were put forward.⑶ Visual positioning model was built based on the visual positioning system platform. Meanwhile, the relationships between different coordinate system in the model were analyzed. On the basis of the above model, the 3D reconstruction method of monocular stereo vision based on the robot end effector was proposed. Then analyze the parameters of inside and outside the camera in the visual positioning andthe methods of determining the above parameters. Finally, extract the solder joint center pixel coordinates needed in the process of 3D reconstruction through image processing.⑷ According to the built visual platform and the proposed visual positioning method, the system software design was completed with VC++ 6.0development environment under the Windows operating system. Lastly, the accuracy and feasibility of the system was verified with the relevant experiment and the visual positioning system met the design requirements.