Study On Machine Vision Technology In Feature Recognition And Inspection Of Stamping Parts

Machine vision technology,which involves image processing,pattern recognition,precision measurement and other technologies,is widely used in industrial production for various characteristics,such as non-contact,flexible positioning,short sampling period,large amount of information,low cost,good stability,high precision and intelligent etc.It is difficult to guarantee the quality and the rate of good products of stamping parts accurately,because of the fluctuation of material properties of stamping parts,the change of working conditions,the repeated handling of workpieces and other factors.The traditional inspection technology of shape size and appearance defect have many problems,such as the low efficiency,the high cost,the low level of complete inspection,the low rate of good products,and difficulties in meeting the problems of mass operation.The application of machine vision technology to stamping pipeline has the advantages of the high efficiency,the high speed,the superior real-time,the high intelligent integration,etc.,which can solve effectively the above problems.Based on the subject that is about precision assembly robot,the paper focuses on the feature recognition technology and visual inspection technology of stamping parts.The hardware vision system of subject is a rotary platform system that has been built after a detailed design and evaluation,which meets the requirements of static performance testing and dynamic repetitions testing of stamping parts under trigger state.Considering the accuracy of the test results,camera calibration is completed combined with the dot calibration board.The software adopts the image processing software HALCON to put forward a method of oriented appearance recognition and visual inspection,and to verify the rationality of the algorithm through experiments.The method is based on main technologies of blob feature analysis,geometric positioning,affine transformation,caliper measurement and edge detection,etc.In the meantime,ageneral visual detection system has been developed,which enables off-line and on-line inspection about target images,as well as has the functions of manual and automatic testing.The main research work and innovation points that concerning the appearance of the target product stamping parts and testing indicators,are as follows:(1)The present situation of stamping parts inspection technology is studied.The traditional stamping parts detection technology and visual inspection technology are compared and analyzed to demonstrate the feasibility of visual inspection technology.Understanding deeply the development and application of visual detection technology by describing its development status at home and abroad in detail.Aiming at the application of visual detection technology in stamping pipeline,this paper introduces the vision system,detection process and the related theory of image processing algorithm in detail.(2)To complete the hardware selection and construction of the experimental platform,we make an analysis of the relevant features of stamping parts and identifying the features based on the accuracy and functional requirements of the test through detailed calculation and comprehensive comparison.To finish the image acquisition on the debugged platform,and to complete the camera calibration experiment for ensuring the accuracy.(3)In this paper,the relevant algorithms of blob analysis for small stamping parts are studied,and the methods of feature recognition are proposed.The principle of various geometric positioning algorithms is studied,and the shape-based template geometric localization algorithm is proposed.The differences between single-template and multi-template matching processes and results are also analyzed.Taking a simple stamping connector as an example,the algorithm of affine transformation is studied,and a single template and multi-template geometric transformation algorithm that based on the image following and the region following are proposed.Taking universal connectors as an example,the two-dimensional measurement algorithms for four different characteristics of connectors are studied,and a new algorithm for fitting straight lines and circles are proposed and the accuracy of the measurement results is proved by experiments.(4)Based on the integrated environment of Visual studio 2015,an universal testing system interface based on C# application is developed.The communication function of the software is realized by using the laptop and serial port assistant.The detection system contains the main algorithms of image processing,which can realize many basic functions such as image acquisition,geometric positioning,fitting straight line and circle,measuring and so on.In the state of communication,two modes of manual detection and automatic detection can be completed.