基于机器视觉的IC芯片焊点定位检测与误差分析

发布时间：2018-06-12 09:13

本文选题：引线键合 + 视觉定位　；参考：《西安工业大学》2015年硕士论文

【摘要】：IC(集成电路)产业是衡量一个国家综合实力的重要指标,是转变发展模式、提升工业水平的重要依托。而集成电路封装技术是IC芯片成为具有实用价值的微电子器件的桥梁,因此研究作为集成电路封装技术中最常用的引线键合技术具有重要意义。机器视觉技术以其非接触性、高效率、高精度等优点在自动化生产中发挥越来越重要的作用。同时为了满足引线键合技术高集成化、高精度化、高细密化的的要求,本论文将机器视觉技术引入引线键合中,以面向引线键合的视觉定位系统为研究对象,根据引线键合对视觉系统的要求搭建实验平台,结合机器视觉技术和图像处理技术设计软件定位算法。本文研究的主要内容如下：首先,对引线键合视觉定位系统进行整体设计,包括实验平台设计和软件算法设计,实验平台设计部分包含硬件选型、光学成像系统的设计等内容；软件算法设计部分包含角点检测、修正量计算等内容。并且着重对整个定位系统中出现的误差进行了分析,具体包括：图像获取时的镜头畸变误差、芯片制造和贴片误差、软件定位算法误差,通过对误差的分析,从而对整个视觉定位算法的优劣性进行准确评估。其次,针对系统的实验平台部分,对获取图像的摄像机镜头进行误差的评定算法设计。该算法采用了传统的摄像机标定算法,以平面圆点阵列型靶标作为研究对象,通过获取的单帧模板图像可求得摄像机的内外参数,同时计算得到图像的镜头畸变误差,并以此作为图像畸变矫正是否进行的依据。经仿真实验验证,本算法计算精度和效率均较高,同时需要强调的是该算法具有较强的普遍适用性。最后,针对裸芯片、基座图像的特点,提出一种基于matlab的软件平台对系统进行定位算法设计,实时的校正在引线键合过程中由贴片和芯片制造产生的误差。通过对裸芯片、基座图像边缘的特点进行分析,本文采用基于几何中心的算法提取出裸芯片、基座图像的虚拟角点,得到二者的实际位置坐标,从而得到二者位移和角度的修正量,并将其结果反馈到键合头,可完成高精度、高效率的键合。经过仿真实验验证,该算法可以实现键合过程中裸芯片、基座位置偏差的实时校正,并且定位精度满足引线键合的工艺要求。综上所述,本论文对引线键合视觉定位系统进行了深入的研究,对所设计的算法均经过仿真实验验证,实验结果表明本文提出的算法均具有较高的检测精度和识别率,达到了比较好的效果。本论文取得的研究成果对于机器视觉定位检测技术、图像畸变矫正技术、虚拟角点检测技术均具有一定的理论价值和实用价值。
[Abstract]:IC (integrated circuit) industry is an important index to measure a country's comprehensive strength, it is an important support to change the development mode and improve the industrial level. The integrated circuit packaging technology is the bridge of IC chip to become the microelectronic device with practical value, so it is of great significance to study the lead bonding technology which is the most commonly used in the integrated circuit packaging technology. Machine vision technology plays a more and more important role in automatic production because of its advantages of non-contact, high efficiency and high precision. At the same time, in order to meet the requirements of high integration, high precision and high fineness of wire bonding technology, this paper introduces machine vision technology into lead bonding, and takes the lead bonding oriented vision positioning system as the research object. An experimental platform is built according to the requirements of lead bonding for visual system. Combined with machine vision technology and image processing technology, software positioning algorithm is designed. The main contents of this paper are as follows: first, the overall design of the lead bond vision positioning system, including the experimental platform design and software algorithm design, the experimental platform design part includes hardware selection, optical imaging system design content; The software algorithm design includes corner detection, correction calculation and so on. The errors in the whole positioning system are analyzed, including lens distortion error, chip manufacturing and patch error, software positioning algorithm error. Therefore, the advantages and disadvantages of the whole vision location algorithm are evaluated accurately. Secondly, the error evaluation algorithm of the camera lens is designed for the experimental platform of the system. This algorithm adopts the traditional camera calibration algorithm and takes the planar dot array target as the research object. The internal and external parameters of the camera can be obtained by the obtained single frame template image, and the lens distortion error of the obtained image can be calculated at the same time. And this is used as the basis of image distortion correction. The simulation results show that the algorithm is accurate and efficient, and it should be emphasized that the algorithm is widely applicable. Finally, according to the characteristics of bare chip and pedestal image, a software platform based on matlab is proposed to design the positioning algorithm of the system, and to correct the errors produced by the chip and the chip in the process of lead bonding in real time. By analyzing the characteristics of the bare chip and the edge of the base image, the algorithm based on geometric center is used to extract the virtual corner of the bare chip and the image of the base, and the actual position coordinates of the two are obtained. The modified displacement and angle can be obtained, and the results can be fed back to the bonding head, which can achieve high precision and high efficiency bonding. The simulation results show that the algorithm can realize the real-time correction of the position deviation of the bare chip and the base during the bonding process, and the positioning accuracy can meet the technological requirements of the lead bonding. To sum up, this paper has carried on the thorough research to the lead bond vision localization system, all the designed algorithms have been verified by the simulation experiment, the experimental results show that the proposed algorithm has a high detection accuracy and recognition rate. Good results have been achieved. The research results obtained in this paper have certain theoretical and practical value for machine vision positioning detection technology, image distortion correction technology, virtual corner detection technology.
【学位授予单位】：西安工业大学
【学位级别】：硕士
【学位授予年份】：2015
【分类号】：TN405;TP391.41

【参考文献】