手机电池的尺寸测量和外观缺陷检测

发布时间：2018-03-02 21:39

本文选题：尺寸测量　切入点：相机标定　出处：《哈尔滨工业大学》2016年硕士论文　论文类型：学位论文

【摘要】：目前,工业领域中对手机电池的尺寸测量和外观缺陷检测的精度要求越来越高,目前在手机电池制造中,尺寸测量和表面缺陷检测都是人工检测,比较耗时耗工,很难达到检测的要求和各种标准。随着机器视觉技术的不断成熟,用于手机电池的自动化检测系统也发展起来,基于机器视觉的检测技术是尺寸测量和外观缺陷检测领域中的先进技术。本文采用机器视觉的方式,对手机电池的尺寸测量和外观缺陷检测进行了研究。对于手机电池的尺寸测量,首先根据测量的精度要达到0.02毫米,选取相机、光源和镜头,搭建出硬件检测平台,用张正友方法实现相机标定,对采集到的图片进行滤波、边缘检测等预处理。设计出一种基于相关系数进行改进的复合匹配算法,实现测量过程中的图像匹配,再对匹配之后的图像进行Hough变换,拟合出直线,根据直线的坐标信息,求出测量的结果。对于手机电池的外观缺陷检测,提出基于仿射变换的图像配准方法,采用差分法实现待检测图像的缺陷判断,对检测到的缺陷图像进行二值化、形态学处理、Blob分析、连通性分析,拟合出检测到的缺陷的最小外接矩形,从而得到缺陷的大小、个数和位置等信息,检测出表面缺陷。最后,用本文所研究的尺寸测量的算法对手机电池的尺寸进行测量,获得了多组数据,分析实验数据,长度测量的标准偏差小于0.01毫米,角度测量的标准偏差小于0.05°,算法精度高,符合工业生产的要求,已经成功应用于工业生产线。采用手机电池的多幅图像对缺陷检测的算法进行验证,计算出缺陷检测的准确率为97.33%。从而验证本文的缺陷检测算法能有效的检测出手机电池的表面污渍、表面白点和图案漏印的缺陷信息,而且准确率较高。
[Abstract]:At present, in the field of industry, the precision of cell phone battery size measurement and appearance defect detection is higher and higher. At present, in the cell phone battery manufacturing, dimension measurement and surface defect detection are both manual detection, which is time-consuming and labor consuming. It is very difficult to meet the requirements of testing and various standards. With the development of machine vision technology, the automatic detection system for mobile phone batteries has also developed. The detection technology based on machine vision is an advanced technology in the field of dimension measurement and appearance defect detection. The dimension measurement and appearance defect detection of mobile phone battery are studied. Firstly, according to the accuracy of measurement, the camera, light source and lens are selected to build a hardware detection platform. The method of Zhang Zhengyou is used to calibrate the camera, filter the collected image, detect the edge and so on. An improved composite matching algorithm based on the correlation coefficient is designed to realize the image matching in the measurement process. Then the matched image is transformed by Hough, and the straight line is fitted out. According to the coordinate information of the line, the measurement results are obtained. For the appearance defect detection of mobile phone battery, an image registration method based on affine transformation is proposed. The defect judgment of the detected image is realized by using the difference method, the detected defect image is binarized, the morphological processing is Blob analysis, the connectivity analysis, the minimum external rectangle of the detected defect is fitted, and the size of the defect is obtained. Finally, the size measurement algorithm studied in this paper is used to measure the size of the cell phone battery, and many sets of data are obtained, and the experimental data are analyzed. The standard deviation of length measurement is less than 0.01 mm. The standard deviation of angle measurement is less than 0.05 掳, the accuracy of the algorithm is high, and it meets the requirements of industrial production. It has been successfully applied to the industrial production line. The defect detection algorithm is verified by using multiple images of mobile phone battery. The accuracy of defect detection is 97.33. It is proved that the defect detection algorithm in this paper can effectively detect the defect information of the cell phone battery, such as surface stain, white spot and pattern missing print, and the accuracy is high.
【学位授予单位】：哈尔滨工业大学
【学位级别】：硕士
【学位授予年份】：2016
【分类号】：TN929.53;TM912;TP391.41

【相似文献】