基于背景连通域的印刷线路板缺陷定位及识别

发布时间：2018-06-26 06:23

本文选题：印刷线路板 + Gerber文件　；参考：《南昌航空大学》2015年硕士论文

【摘要】：印刷线路板是现代电子信息产业中最重要的元件之一,朝着高密度、微型化发展,其制作过程中若未及时对缺陷定位识别,将导致整块PCB报废。因此,PCB的缺陷定位与识别方法研究对提高其制作质量具有重要意义。本文对线路板标准图像和待检测图像的配准技术以及线路板缺陷的定位与识别技术展开了深入的研究,主要研究工作和成果如下:(1)研究了Gerber文件生成线路板标准图像及待测图像的采集和预处理方法。首先采用设计绘制线路板的过程中产生Gerber文件并将其解析,生成线路板标准图像;其次使用CCD相机采集线路板实物图作为检测对象;然后采用快速中值滤波方法去除图像采集时引入的高斯噪声与椒盐噪声;最后对图像进行阈值分割,将线路板焊盘、线条等主要特征图像与背景分离。(2)研究了线路板标准图像与待测图像配准方法。方法首先通过解析Gerber文件获取标准图像特征点参数;其次根据标准图像特征点参数计算出待测图像特征点搜索范围,并通过连通域法找出定位圆位置、通过Canny算子边缘检测获得定位圆边界点的坐标,再使用最小二乘拟合办法计算出精确的待测图像特征点参数;最后通过相似变换对待测图像进行角度和大小的校正,再将变换后的待测图像按禁止布线框位置裁剪,完成标准图像与待测图像的精确配准。实验结果表明,该方法能对线路板标准图像与待测图像做精度较高的配准。(3)研究了基于背景连通域的线路板缺陷定位及识别方法。方法第一步对线路板缺陷进行定位,首先,将实现配准后的标准图像与待测图像进行差影算法异或运算,获得缺陷初步定位图;其次,使用形态滤波开运算滤除由于配准细微误差留下的重影,并设定阈值滤除初步定位图中杂质;最后采用轴向包围盒对缺陷图像进行标记,完成缺陷定位。方法第二步对线路板缺陷进行识别。首先,将所标记出的缺陷一一编号,以利于识别过程中缺陷信息的对应保存;其次,通过基于铜料检测的方法将缺陷进行一次分类,分为多铜缺陷与少铜缺陷;最后,使用基于背景连通域的办法进行缺陷识别,通过分别对每处缺陷区域标准图像和待测图像的前景连通域及背景连通域数量进行比较,实现对线路板常见缺陷类型的识别。实验结果表明,该方法能对常见缺陷类型有效定位及识别。
[Abstract]:Printed circuit board (PCB) is one of the most important components in modern electronic information industry. It is developing towards high density and miniaturization. If the defects are not identified in time in the manufacturing process, the whole PCB will be scrapped. Therefore, the research of PCB defect location and identification is of great significance to improve the quality of PCB fabrication. In this paper, the registration technology of PCB standard image and the image to be detected, as well as the localization and recognition technology of PCB defects are deeply studied. The main work and results are as follows: (1) the methods of collecting and preprocessing the standard image and the image to be tested by Gerber file are studied. In the process of drawing circuit board, Gerber file is generated and parsed to generate the standard image of circuit board, and then the physical diagram of circuit board is collected by CCD camera as the object of detection. Then the fast median filtering method is used to remove the Gao Si noise and salt and pepper noise introduced in the image acquisition. Finally, the threshold value of the image is segmented, and the circuit board solder is used. Lines and other main feature images are separated from the background. (2) the registration method between the standard image and the image to be tested is studied. Methods the parameters of standard image feature points were obtained by analyzing Gerber file, and then the search range of feature points was calculated according to the parameters of standard image feature points, and the location of the circle was found by means of the connected domain method. The coordinates of the circular boundary point are obtained by Canny operator edge detection, and the parameters of the feature point of the image under test are calculated by using the least square fitting method. Finally, the angle and size of the measured image are corrected by the similarity transformation. Then the transformed image is clipped according to the position of the forbidden cabling frame to complete the accurate registration of the standard image and the image to be tested. The experimental results show that the proposed method can be used to register the standard image and the image to be tested. (3) the defect location and recognition method based on the background connected domain is studied. Methods the first step is to locate the defects of the circuit board. Firstly, the standard image and the image to be tested are computed by the difference or algorithm, and the initial localization map of the defect is obtained. Morphological filtering is used to filter the shadow left by minor registration errors, and the threshold is set to filter the impurities in the initial location map. Finally, the defect image is labeled with an axial bounding box to complete the defect location. Methods the second step was to identify the defect of circuit board. First, the defects are numbered one by one, so as to facilitate the corresponding preservation of defect information in the process of identification. Secondly, the defects are classified into more copper defects and fewer copper defects through a method based on copper material detection. The method based on background connected domain is used to identify defects. By comparing the number of foreground connected domain and background connected domain of each defect region standard image and the image to be tested, the common defect types of PCB can be identified. Experimental results show that this method can effectively locate and identify common defect types.
【学位授予单位】：南昌航空大学
【学位级别】：硕士
【学位授予年份】：2015
【分类号】：TN41;TP391.41

【参考文献】