【作者】 杨洁；

【导师】 侯开虎；

【作者基本信息】 昆明理工大学 ， 工业工程（专业学位）， 2017， 硕士

【摘要】 产业的调整及不断升级对产品的整体质量提出了更高的要求,但在整体的印制电路板(Printed Circuit Board,PCB)的制造中,一般水平的PCB过剩,而高质量高附加值的PCB却供不应求。常见的PCB的缺陷检测方法一般都是人工事后检测,该方法会出现漏检、错检的情况,而且不能及时发现并解决问题,因此,一款精度有保障的PCB质量在线检测设备是必须的。本文以PCB制造企业WK为背景,拟研究PCB裸板运用图像处理的在线检测过程。本文借鉴国内外相关行业产品,相关算法以及行业经验,设计了一套基于图像识别的在线检测系统,以图像识别为手段,以搭建的图像采集硬件系统为基础,对PCB裸板进行在线检测。目的是在采集高质量的图片基础上,进行最准确的图像识别并判断PCB裸板是否符合加工要求。本系统主要包括两大部分:硬件部分为该系统的基础支持,主要由图像采集系统,运动控制系统,照明系统三部分组成,其中图像采集系统采用的是CCD(工业照相机)的线阵照相机;软件系统为本系统的核心部分,该部分的主要功能是对硬件系统进行控制,在采集到图像之后,对图像进行处理并判断。本文涉及到的主要算法有,摄像头自动对焦算法、图像拼接算法、图像预处理算法,圆孔图像识别算法。结合系统硬件本身的条件选择最适合本系统的算法构建了一套基于图像识别的PCB裸板在线监测系统平台。通过对系统主要功能的测试,实现了摄像机的自动对焦,对算法进行优化选择之后的图像拼接准确率得到了提高,图像识别的效率准确度得到了保障。采用高精度图像定位,定位功能更加准确,通过对匹配度的数据显示可以直观的判断加工检测件是否合格。与市场上其他的PCB检测设备进行对比,该系统硬件成本较小,采集图像的效果相对稳定,软件界面简单、操作方便、便于一线工人操作使用,更加的符合我国市场的实际需求。更多还原

【Abstract】 The continuous adjustment and upgrading of industries put forward higher requirements for the overall quality of the product.But the phenomenon is commonly seen in the manufacture of overall Printed Circuit Board(Printed Circuit Board,PCB)that the PCBs of general level are surplus,while these owning high quality and are in short supply.As the common defect methods of PCB,the artificial detection after defect occurring make it possible that the situation of missing or wrong detection and cannot solve the problem existed in time,so a precision-guaranteed quality online testing equipment for PCB is necessary.The online detection process based on the image recognition of PCB bare board in the background of WK company was researched in the thesis.The on-line detection system based on the image recognition was designed according to the experience of related products and algorithms at home and abroad in this thesis to detect the PCB bare board on line by means of image recognizing and building the hardware system for image acquisition.The purpose was to recognize most accurately the images and judge whether the PCB bare board meet the processing requirements based on the high-quality images collected.The system is consisted of two parts:the hardware part of the system is as the basic support,composed of the image acquisition system,which utilizes the linear array camera of CCD(industrial camera),motion control system and lighting system The hardware part is controlled by the other core part,the software system,the main function of which is to processing and judging these images after collecting them.The camera auto-focus algorithm,image splicing algorithm,image preprocessing algorithm and round hole image recognition algorithm are revolved as the main algorithms in this thesis.On-line monitoring system platform of the PCB bare board based on image recognition is build by choosing the most suitable algorithm combined with the conditions of the hardware itself.The auto focus of the camera is realized Through the testing of the main functions of the system.The accuracy of the image splicing is improved and the efficiency of the image recognition get guaranteed after the algorithm being optimized.The adoption of high precision image positioning makes positioning function more accurate.Whether the products to be tested are qualified can be intuitively determined through the matching data displayed.Compared with other PCB detection equipment on the market,the hardware cost of the system is relatively small and the effect of collecting images is relatively stable.And the software interface is simple and its operation is convenient for the first-line workers to operate,which meets greatly the actual demand of market in China.更多还原