Research On The Vision Alignment System Of High Precision Electronic Packaging Machine

Electronic packaging is a very important process in semiconductor manufacturing,because the quality of the packaging has direct impact on the performance of semiconductor products.The mechanical and the inductive alignment system have been unable to meet the current need of the industry in precision and stability,and the machine vision alignment system gradually replaces the traditional alignment system with the advantage of high precision and high robustness.At present,most vision alignment system is composed of some dedicated hardware,and its software development is mostly based on professional visual development software,which often makes the system more expensive.Achieving visual alignment system with general-purpose hardware and software,not only can reduce system cost,but also can reduce the system build cycle.At the same time,traditional vision alignment algorithm also has some shortcomings,which need to be improved.In this paper,the vision alignment system for electronic packaging is studied,and the structure of vision alignment system is analyzed,and the experimental platform of vision alignment system with general hardware is built,then the visual alignment algorithm are analyzed and improved,finally the vision alignment system is verified by experiments.The main work is as follows:(1)Hardware composition and function of vision alignment system are studied.The paper makes a deep research on its machine vision system.Considering the narrow space of double stage,a special optical path acquisition system is used.At the same time,parameter selection of each hardware of the system is studied,and the experiment platform of visual alignment system is built with general hardware.(2)The vision alignment algorithm is analyzed.A variety of image processing and analysis algorithm of vision alignment algorithm is studied,and an improved Lee filtering algorithm is proposed in the image preprocessing algorithm;and the mapping algorithm is improved,and a location algorithm based on Hough transform and the least square method is used.(3)A set of vision alignment system software is developed by using OpenCV,MFC and other common software development kits.It contains all the image processing and analysis algorithms and motion control algorithms.(4)The performance of the vision alignment system is verified.The experimental results show that the alignment error of the vision alignment system is within ±8?m,and the rotation alignment error is within ±0.08°.At the same time,the experimental results are analyzed to find the place where the experiment system needs to be improved.