印制电路板制造中缺陷检测的研究

发布时间：2018-01-28 21:59

本文关键词： 印制电路板圆心检测图像拼接缺陷检测　出处：《大连理工大学》2015年硕士论文　论文类型：学位论文

【摘要】：印制电路板(PCB)是为当今电子产品提供电气连接的关键互连组件,为了提高电子产品的稳定性,就需要品质优良的印制电路板。所以,在印制电路板生产过程中,缺陷检测是不可或缺的环节。而传统的人工的检查在大量的生产中会因为误判率高、速度慢、漏检率高等缺陷导致生产停滞,因此,由于自动光学检测技术[1]在速度、精度与费用等方面更有优势,所以其可以替代人工检查。本课题主要探讨自动光学检测技术[1]在印制电路板制造中的应用。基于图像处理技术和机器视觉理论,本文主要研究了系统各环节算法,使用Matlab编译环境对获取的图像进行分析比对。本课题首先研究对印制电路板内层板首个工序,基板曝光中全自动曝光机对胶片同基板的靶标,计算其类同心圆圆心的位置关系。然后运用自动光学检测机(AOI)拍摄图片,由于拍摄的图片为连续并具有部分重复特征的图片,需要运用图片拼接技术对其进行拼接。最后对内层板工序中的断路、针孔等问题进行提取和鉴别,从而达到较好的质量检测和控制效果。本课题工作主要可以概括为以下3个方面：(1)对内层板制造中全自动曝光机靶点定位图片进行圆心检测,通过图像灰度化,图像二值化,检测图像中的边缘点,运用霍夫变换精确检测两圆圆心坐标,计算出内层板孔圆心同胶片定位靶点圆心之间的距离。(2)对内存板自动光学检测机(AOI)行扫描后CCD拍照的照片进行拼接,研究了SIFT算子的特征,其对形状类似的物体也能产生大量的特征向量。实验发现在降低匹配精度的情况下,不但能很好的找到相同的特征点,而且还大大提高了工作效率。利用匹配到的特征点,基于最小二乘法对图片进行拼接。(3)对于PCB缺陷检测,本位主要运用相关参考法。首先采集标准PCB图像并灰度化,然后再进行图像预处理；接着同样采集并灰度化待测PCB,最后对处理后的二图做相减操作,保留标准图与待测图的对比图,经过形态学处理后,分析缺陷特征并记录。
[Abstract]:Printed circuit board (PCB) is a key interconnection component to provide electrical connection for today's electronic products. In order to improve the stability of electronic products, high quality printed circuit boards are needed. In the production process of printed circuit board, defect detection is an indispensable link, and the traditional manual inspection in a large number of production will be due to the high rate of misjudgment, slow speed, high rate of missed detection and other defects leading to production stagnation. Automatic optical detection technology. [1) it has more advantages in speed, precision and cost, so it can replace manual inspection. This paper mainly discusses the technology of automatic optical inspection. [Based on the image processing technology and the theory of machine vision, this paper mainly studies the algorithm of every link of the system. Using Matlab compiler environment to analyze and compare the obtained images. Firstly, this paper studies the target of the first working procedure of the printed circuit board inner laminate, the automatic exposure machine in the substrate exposure to the same substrate. The position relation of the concentric circle center is calculated, and then the image is taken by the automatic optical detector (AOI), because the image is continuous and has the characteristic of partial repetition. It is necessary to use the technology of picture stitching. Finally, the problems such as broken road and pinhole in the inner laminate process are extracted and identified. In order to achieve a better quality detection and control effect. This work can be summarized as the following three aspects: 1) the automatic exposure machine target location pictures in the manufacture of inner laminates to detect the center of the circle. The edge points in the image are detected by image grayscale and image binarization, and the coordinates of two circular centers are accurately detected by Hough transform. The distance between the hole center of the inner laminate and the center of the target point of the film was calculated. (2) the CCD photographs taken by the memory plate automatic optical detector were stitched after scanning. We study the features of SIFT operators, which can also produce a large number of feature vectors for objects with similar shape. Experiments show that in the case of reducing the matching accuracy, not only the same feature points can be found very well. It also greatly improves the working efficiency. Using the matching feature points, the paper uses the least square method to mosaic the image. 3) for the PCB defect detection. First, the standard PCB images are collected and grayscale, and then the image preprocessing is carried out. Then the PCBs are also collected and grayscale to be tested. Finally, the two images are subtracted, and the contrast between the standard map and the new map is retained. After morphological processing, the defect features are analyzed and recorded.
【学位授予单位】：大连理工大学
【学位级别】：硕士
【学位授予年份】：2015
【分类号】：TN41;TP391.41

【相似文献】