基于微观图像的PCB板表观缺陷检测算法研究

发布时间：2018-10-29 19:30

【摘要】：印刷线路板(PCB)是各种电子元器件的载体,随着电子设备发展而向高密度、高集成、轻薄化、易穿戴的方向演进,使得PCB表观缺陷检测的自动化成为业界重要研究领域。本文以自动光学检测(AOI)为基础,以PCB板表观缺陷检测及其速度和准确率提高为目标,提出了一种基于微观图像的PCB板表观检测算法,可精准快速地检测出高密度PCB板的表观缺陷。本文研发的检测算法首先改进了光源模块,使得采集图像光照强度更均匀。其次提出了针对性降噪的图像预处理方案,综合应用高斯滤波、中值滤波和小波滤波来消除各种噪声干扰,突出图像的边缘特征。其三基于SIFT算法和Harris角点,提出了一种快速实现微观图像的图像拼接算法及其结合实际需求的优化,实时性更好。其四通过PCB板存在表观缺陷的分类,结合形态学而提出了缺陷检测算法,能精准检测并识别出不同表观缺陷。从而,在实现微观局部的图像拼接成完整高密度PCB板图像基础上,解决了高密度、高集成PCB板检测难、效率低的行业难题,适应PCB实际生产应用需求,为PCB发展奠定了良好基础,对PCB产业新应用具有良好的理论指导和工程应用价值。
[Abstract]:Printed circuit board (PCB) (PCB) is the carrier of various electronic components. With the development of electronic devices, it has evolved to the direction of high density, high integration, thinning and wearability, which makes the automation of PCB apparent defect detection become an important research field in the industry. Based on automatic optical detection (AOI) and aiming at improving the speed and accuracy of PCB plate apparent defect detection, this paper presents a microscopic image based PCB plate apparent detection algorithm. The apparent defects of high density PCB plate can be detected accurately and quickly. The detection algorithm developed in this paper firstly improves the light source module to make the light intensity of the collected image more uniform. Secondly, an image preprocessing scheme is proposed, which combines Gao Si filter, median filter and wavelet filter to eliminate all kinds of noise interference and highlight the edge features of the image. Thirdly, based on the SIFT algorithm and Harris corner, a fast image stitching algorithm is proposed, which combines with the actual requirements, and the real-time performance is better. Fourthly, through the classification of apparent defects in PCB board and combining with morphology, a defect detection algorithm is proposed, which can accurately detect and identify different apparent defects. Therefore, on the basis of realizing micro local image splicing to complete high density PCB board image, the industry problem of high density and high integration PCB board is solved, which is difficult to detect and low efficiency, and adapts to the practical production and application demand of PCB. It lays a good foundation for the development of PCB and has good theoretical guidance and engineering application value for the new application of PCB industry.
【学位授予单位】：杭州电子科技大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TN41;TP391.41

【参考文献】