华南理工大学学报(自然科学版) ›› 2010, Vol. 38 ›› Issue (1): 65-69,86.doi: 10.3969/j.issn.1000-565X.2010.01.013

• 电子、通信与自动控制 • 上一篇    下一篇

基于视觉的贴片元件检测算法

张舞杰1 李迪叶峰2   

  1. 1.华南理工大学 自动化科学与工程学院, 广东 广州 510640; 2.华南理工大学 机械工程与汽车学院, 广东 广州 510640
  • 收稿日期:2008-12-30 修回日期:2009-06-28 出版日期:2010-01-25 发布日期:2010-01-25
  • 通信作者: 张舞杰(1970-),男,博士后,主要从事图像处理、模式识别、过程监控、嵌入式装备控制研究. E-mail:zwjllhtt@scut.edu.cn
  • 作者简介:张舞杰(1970-),男,博士后,主要从事图像处理、模式识别、过程监控、嵌入式装备控制研究.
  • 基金资助:

    中国博士后科学基金资助项目(20070420784);广东省自然科学基金博士科研启动项目

Vision-Based Inspection Algorithm for Chip Components

Zhang Wu-jieLi DiYe Feng2   

  1. 1. School of Automation Science and Engineering, South China University of Technology, Guangzhou 510640, Guangdong, China ; 2. School of Mechanical and Automotive Engineering, South China University of Technology, Guangzhou 510640, Guangdong, China
  • Received:2008-12-30 Revised:2009-06-28 Online:2010-01-25 Published:2010-01-25
  • Contact: 张舞杰(1970-),男,博士后,主要从事图像处理、模式识别、过程监控、嵌入式装备控制研究. E-mail:zwjllhtt@scut.edu.cn
  • About author:张舞杰(1970-),男,博士后,主要从事图像处理、模式识别、过程监控、嵌入式装备控制研究.
  • Supported by:

    中国博士后科学基金资助项目(20070420784);广东省自然科学基金博士科研启动项目

摘要: 为了实现贴片元件的自动检测,提出了一种基于视觉的贴片元件几何特征参数检测方法.首先采用最大外接矩形法实现元件的粗定位及确定边缘的分割点,并采用Canny和Zernike矩边缘检测算子实现边缘的精确定位.然后,利用分割点将边缘分割成4部分,分别进行直线和圆弧拟合,得到其精确值.同时,利用快速傅里叶变换后的图像特征,实现端面图像中条纹方向的判定.实验中测得亚像素边缘点的定位精度为0.03像素,直线拟合精度为0.03像素,圆弧拟合精度为0.05像素,端面条纹判断的准确率为100%.实验结果表明:文中提出的检测方法能很好地满足贴片元件自动视觉检测稳定可靠、精度高及实时性强的要求.

关键词: 贴片元件, 亚像素, 视觉检测, 快速傅里叶变换, 边缘检测, 拟合

Abstract:

In order to automatically inspect chip components, a vision-based method is proposed to calculate the geometrical parameters of the components. In this method, first, the coarse location of chip components and the edge point sorting are realized by means of the maximum external rectangle method, and the precise location of the edge is implemented by using the Canny operator and the Zernike moment operator. Next, the edge points are sorted into 4 parts according to sorting points, which are then fitted respectively via line and arc fittings to obtain the corresponding accurate values. Moreover, the stripe direction of the transverse image of chip components is correctly judged according to the image characteristics obtained via the fast Fourier transform (FFT). Finally, an experiment is carried out, with a subpixel location precision of 0.03 pixel, a line fitting precision of O. 03 pixel, an arc fitting precision of 0. 05 pixel and a stripe direction accuracy of the transverse image of 100% being obtained. The results indicate that the proposed inspection method is of strong stability, high precision and excellent real-time perfor- mance, which is helpful in the automatic Vision-based inspection of chip components.

Key words: chip component, subpixel, vision inspection, fast Fourier transform, edge detection, fitting