The Research On Welding Components And Solder Joint Recognition Algorithm Of Automobile Door Panel

In the actual production process of automobile door panels,the types of automobile door panels change with the factory production needs,which will inevitably lead to the corresponding changes of the weldment types and locations of the solder joints on the automobile door panel,however,traditional welding robot can not automatically detect these uncertainties,as a result,the welding quality and efficiency of the automobile door panel often fail to meet the qualifications.Therefore,in order to assist the welding operation of the welding robot and improve the efficiency of the welding process of the automobile door panel,this paper studies the welding components and solder joint recognition algorithm based on the traditional image processing algorithm and the convolutional neural network.The main research contents of this paper are as follows.Firstly,the research background of this paper is introduced.The significance of this paper is expounded in view of the problems existing in the traditional automobile door panel welding production process.Subsequently,the problem of welding components and solder joint identification of automobile door panels was analyzed,including the process of automobile door panel welding technology and automobile door panel welding production line,the objectives and tasks of welding parts and solder joint identification were clarified,and the corresponding technical routes were given.Second,the welding components and solder joint recognition algorithm based on the traditional image processing was studied.According to the welding characteristics of the car door panel,this paper proposes a welding componets identification method combining Canny edge detection and SSAD(sequential similarity detection)algorithm.In the process of solder joint identification,the image of the automobile door panel is segmented by the identified welding components,and the position and number of solder joints in the segmented image are identified by the Hough transform algorithm,which greatly improves the accuracy and efficiency of solder joint identification.Then,the welding components and solder joint recognition algorithm based on convolutional neural network was studied.Based on the analysis of the welding components and solder joint identification targets and the shortcomings of the YOLO v1 algorithm,this paper uses the improved YOLO v3 algorithm for soldering components and solder joint identification.The improved version of the YOLO v3 algorithm can significantly improve the recognition accuracy of small targets such as solder joints due to the introduction of the Residual structure in the ResNet network.and the experimental results prove the effectiveness of the improved YOLO v3 algorithm.Finally,the experimental results of the traditional algorithm and the YOLO algorithm are compared and analyzed.The advantages and disadvantages of the two algorithms are compared,and their use conditions in different scenarios are given.