Research On Laser Welding Location And Defect Detection Technology Of Silver Net In Pb-Zn Battery Based On Machine Vision

The welding process of silver net is a necessary process in the manufacturing process of lead zinc battery.At present,the welder still use manual electric iron to weld the net.This welding method not only has problems like inaccurate identification with naked eyes and high error rate,but also requires a high labor intensity and features low efficiency.With the development of increased demand for intelligent equipment in welding industry,welding robot become more and more popular,and applying the machine vision technology to the field of solder joint identification and defect detection is equivalent to the giving welding equipment eyes,therefore it has been an important research point in the field of welding machine.Applying the machine vision technology to the welding process can replace human eyes and achieve high speed and high accuracy of welding spot location.The machine vision technology can also detect the defect of welding spot at the same time,which can achieve the double goals of good quality and high working efficiency.An automatic position welding and welding spot defect detection solution based on machine vision was brought up within this welding process of silver net.The system uses a robot carrying an automatic welding device to weld the spots after identifying them accurately,and the main welding parameters and models of hardwares of the system are determined by the analysis and experiment;adopts the distortion correction technology and robot's hand-eye calibration to complete the accurate positional shift from the image coordinate system to the robot's coordinate system;applies Blob method to analyze the algorithm of the coarse positioning combined with the accurate positioning using shape-based template matching technique,removes the barricades to accurate identification and positioning caused by noise jamming,light interference and deformation of the silver net,finally achieve the goal of highly accurate identification and positioning and high efficiency;an software of upper computer was designed by C++ combined with machine vision library,which includes image processing module,coordinate transformation module,and communication module,and the test shows these modules can meet the functional system requirements;A robot visual positioning experimental platform is built,and the experiment indicates that the accuracy of the localization algorithm can meet the requirements of the technology.In addition,the welding spot defect detection algorithm based on the SVM method and the template matching was designed,which can effectively detected the lack of welding,appearance and positions of welding spot defects,the experimental results show that the recognition rate of the algorithmis higher than that of the ordinary SVM algorithm and has broad application.