基于图像处理技术的AOI系统的研究

发布时间：2018-06-04 02:00

本文选题：印刷电路板 + 图像处理　；参考：《浙江理工大学》2017年硕士论文

【摘要】：在电路板的生产过程中,由于元器件一般都是通过贴片机、插件机或者人工的方式放置到电路板上,故往往会出现一些错误,例如漏件、错件、方向错误等。而目前市面上电路板自动光学检测设备基本上都是针对焊点面的,对于焊点缺陷或者位置固定的贴片元器件缺陷的检测效果较好,但对于元器件面上的直插元器件的检测准确度差、误报率高。因此,对于元器件面的检测大部分都是靠人工检测,其检测过程枯燥重复,非常容易发生漏检和误检。基于以上现状,本文研究的基于图像处理技术的自动光学检测系统,能够实现对电路板元器件面进行缺陷检测。本系统包括图像采集模块、图像处理模块、程序界面模块和数据存储模块四个部分。同时,通过对本系统实验结果的分析,证明了本文所研究的基于图像处理技术的AOI系统具有可行性。本文的研究内容及主要成果如下:(1)图像采集模块的硬件部分主要由VT-EX1400CPS工业相机、VT-LEM0814CB-MP8镜头、箱式无影光源组成;软件部分基于labview语言和VAS视觉获取模块编写。1400万像素的工业相机保证了采集到的图像分辨率高;无影光源保证了电路板上光线照射均匀,元器件阴影降到最低。本模块实现了高质量、低干扰的电路板图像实时采集功能。(2)元器件存在性检测提供了三种检测方法,分别为颜色提取、计算相似度、模板匹配,多种检测方法能够满足绝大多数元器件检测的需求。(3)二极管极性检测算法针对二极管管体位置的不确定性,使用了两种二极管定位的方法,分别是阈值分割方法和查找管脚方法方法;然后通过对二极管管体进行分析,从而得出二极管的极性。(4)电解电容极性检测算法使用改进的霍夫圆检测方法定位电解电容;然后将电解电容带有极性标志的圆环截取下来并展开,通过OTSU方法进行阈值分割,从而判断极性。(5)插座方向检测算法针对插针到两边距离的不同以及插针位置的固定性,使用了一种通过计算插针到插座边缘距离的方法来判断插座方向。
[Abstract]:In the production process of the circuit board, because the components are usually placed on the circuit board by placement machine, plug-in machine or manual way, there are often some mistakes, such as missing parts, wrong direction and so on. At present, circuit board automatic optical inspection equipment on the market is basically aimed at the solder joint surface, and the detection effect of solder joint defect or fixed position patch component defect is better. But the detection accuracy is poor and false alarm rate is high. Therefore, the detection of components surface mostly rely on manual detection, the detection process is boring and repetitive, and it is easy to miss and misdetect. Based on the above situation, the automatic optical detection system based on image processing technology is developed in this paper, which can detect the defects of circuit board components. The system includes four parts: image acquisition module, image processing module, program interface module and data storage module. At the same time, through the analysis of the experimental results of the system, it is proved that the AOI system based on the image processing technology studied in this paper is feasible. The research contents and main results of this paper are as follows: 1) the hardware of the image acquisition module is mainly composed of the VT-LEM0814CB-MP8 lens of VT-EX1400CPS industrial camera and the box type non-shadow light source. The software is based on the labview language and the VAS visual acquisition module. The 14 million pixel industrial camera ensures the high resolution of the collected image, the non-shadow light source ensures that the light on the circuit board is uniform, and the shadow of the components is reduced to the minimum. In this module, the high quality and low interference function of real-time image acquisition of circuit board is realized. It provides three detection methods, namely, color extraction, calculation of similarity, template matching, etc. A variety of detection methods can meet the needs of the majority of components. The diode polarity detection algorithm uses two diode location methods in view of the uncertainty of diode tube position. The method of threshold segmentation and the method of finding pin are used respectively, and then through the analysis of diode tube, the polarity detection algorithm of electrolytic capacitance is obtained, and the improved Hoff circle detection method is used to locate the electrolytic capacitance. Then the ring with polarity mark is cut off and expanded, and then the threshold value is segmented by OTSU method to judge the direction detection algorithm of polarity. 5) aiming at the different distance between the two sides of the pin and the fixing of the pin position, A method is used to determine the direction of the socket by calculating the distance from the pin to the edge of the socket.
【学位授予单位】：浙江理工大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TN41;TP391.41

【参考文献】