Research On Automatic Assembly Measurement And Control Technology Of Small And Medium-sized Transformers Based On Machine Vision

Medium and small transformers have many components and complex assembly process.Traditionally,it is assembled by hand,this method not only has high labor intensity and low efficiency,but also can not guarantee assembly accuracy and the product consistency is poor.In order to solve the problems mentioned above,machine vision is introduced into the assembly process of medium and small transformers.The camera can replace human eye to measure and locate the target parts,and then the robot according to the feedback information from camera to complete the corresponding assembly actions.On this basis,a set of automatic assembly measurement and control system of medium and small transformers based on machine vision is designed,focusing on system parameter calibration and target parts recognition and positioning.Firstly,according to the automatic assembly requirement of medium and small transformers,the problems of parts recognition and positioning and system motion control in assembly process are studied,the control mode of " look and move" and the assembly strategy of " recognition and positioning,parts grasping and parts assembly" are determined.On this basis,the scheme of the automatic assembly and the assembly process are completed.Secondly,system parameter calibration is studied,it discussed from two aspects,which are the visual system and position error.One is about visual system calibration,an optimized hand-eye calibration model based on minimizing the reprojection error is established,and model parameters are optimized by the DFP algorithm of quasi-Newton algorithm,so that the accurate hand-eye conversion matrix can be obtained.Another is about position error calibration,an automatic compensation method of actuator error is proposed.The actuator error can be measured by the experiments of robotic grasping and assembling.The data obtained are substituted into the error model and though several parametric iteration,the optimal error compensation parameters can be obtained and then embedded its into the execution program to complete the correction of position error.Then,according to the characteristics of the original image of the target part,the corresponding algorithm of image recognition and positioning is designed,including feature extraction of image,the establishment of part's standard feature databases,the method based on target feature matching and minimum bounding rectangle algorithm positioning method.Finally,from the two aspects of system hardware constitution and software design,we can explain the specific realization process of automatic assembly measurement and control system for medium and small transformers.Through the actual assembly experiment,it shows that the automatic assembly measurement and control system of medium and small transformers can not only meets the requirements of parts assembly accuracy,but also effectively raises production efficiency and the qualification rate of products.It is suitable for the production process of large quantities of medium and small transformers.