J4 ›› 2014, Vol. 41 ›› Issue (3): 110-115+173.doi: 10.3969/j.issn.1001-2400.2014.03.016

• 研究论文 • 上一篇    下一篇

一种电子元器件组装结果检测方法

姜建国1;王国林1,2;孟宏伟1;李昉1,3;曾奕1,4
  

  1. (1. 西安电子科技大学 计算机学院,陕西 西安  710071;
    2. 中国人民解放军63726部队,宁夏 银川  750001;
    3. 空军工程大学 空管领航学院,陕西 西安  710051;
    4. 中国人民解放军94170部队,陕西 西安  710082)
  • 收稿日期:2013-01-28 出版日期:2014-06-20 发布日期:2014-07-10
  • 作者简介:姜建国(1956- ),男,教授,E-mail: jgjiang@mail.xidian.edu.cn.
  • 基金资助:

    国家部委基础科研计划资助项目(A1120110007)

Detection method for assembling results of electronic components

JIANG Jianguo1;WANG Guolin1,2;MENG Hongwei1;LI Fang1,3;ZENG Yi1,4   

  1. (1. School of Computer Science and Technology, Xidian Univ., Xi'an  710071, China;
    2. No.63726 Unit of PLA, Yinchuan  750001, China;
    3. Air Control & Navigation College, Air Force Engineering University, Xi'an  710051, China;
    4. No.94170 Unit of PLA, Xi'an  710082, China)
  • Received:2013-01-28 Online:2014-06-20 Published:2014-07-10

摘要:

为了更快更准确地检测印刷电路板(PCB)上的元器件是否存在漏焊、误焊或极性倒置等装配错误,提出了一种基于类电磁机制算法的电子元器件组装结果检测方法.该方法利用类电磁机制算法强大的搜索策略和智能的寻优机制,通过随机均匀抽取目标区域坐标点产生初始种群;以改进的归一化相关函数作为目标函数,把图像的匹配同名点看作带电粒子,通过粒子的移动与更新搜索图像的最佳匹配点;引入非均匀变异算子对移动粒子操作进行了改进;给出了该方法的实现步骤和基本流程.实验结果表明,新的检测方法具有更快的收敛速度和更高的识别率,适用于PCB板的检测.

关键词: 电子元器件, 组装结果检测, 类电磁机制算法, 模板匹配, 归一化相关函数

Abstract:

In order to detect the common problems on PCB faster and more accurately, such as leakage welding, false welding or polarity inversion, a new detection method for assembling results of electronic components is proposed, which makes full use of the powerful search strategy and intelligent optimization mechanism of the EM algorithm. In the new algorithm, the initial population is generated by extracting the coordinate points from the feasible domain randomly and uniformly. The improved normalized correlation function is taken as the objective function. The images' matching points with the same name are regarded as charged particles. And the best image matching point is searched by moving and updating particles. Moreover, a non-uniform mutation operator is introduced to improve the operation of moving particles. Finally the implementation steps and basic process of the new method are designed. Experimental results show that the new method has a faster convergence speed and a higher recognition rate, and is suitable for the detection of PCB.

Key words: electronic components, assembling results detection, electromagnetism-like mechanism algorithm, template matching, normalized correlation function

中图分类号: 

  • TP391.4