• 中国计算机学会会刊
  • 中国科技核心期刊
  • 中文核心期刊

J4 ›› 2012, Vol. 34 ›› Issue (6): 88-92.

• 论文 • 上一篇    下一篇

基于纹理特征的磁片表面刀纹缺陷检测

贺延涛1,徐琪2,唐亮1,游晓明3   

  1. (1.复旦大学计算机科学技术学院,上海 201203;2.上海海事大学信息工程学院,上海 201306;
    3.上海工程技术大学电子电气工程学院,上海 201620)
    (1.School of Computer Science,Fudan University,Shanghai 201203;
    2.School of Information Engineering,Shanghai Maritime University,Shanghai 201306;
    3.School of Electrical and Electronic Engineering,
    Shanghai University of Engineering Science,Shanghai 201620,China)
  • 收稿日期:2011-10-16 修回日期:2012-01-05 出版日期:2012-06-25 发布日期:2012-06-25
  • 基金资助:

    上海市自然科学基金资助项目(09ZR1420800)

Detection of Mark Defects on the Disk Surface Based on the Texture Feature

HE Yantao1,XU Qi2,TANG Liang1,YOU Xiaoming3   

  1. (1.School of Computer Science,Fudan University,Shanghai 201203;
    2.School of Information Engineering,Shanghai Maritime University,Shanghai 201306;
    3.School of Electrical and Electronic Engineering,
    Shanghai University of Engineering Science,Shanghai 201620,China)
  • Received:2011-10-16 Revised:2012-01-05 Online:2012-06-25 Published:2012-06-25

摘要:

表面缺陷检测在工业生产中对产品质量可以起到有效的监督控制作用,而目前对磁性材料表面刀纹缺陷检测的方法各自存在自身的局限性,如成本过高、检测速度太慢而不能满足工业生产中实时检测的需求等。为了能够达到实时稳定的检测磁片表面刀纹缺陷的目的,本文借助于计算机工业视觉系统,基于纹理特征,通过图像预处理,采用增强缺陷部分并抑制背景部分的方法,使得各种缺陷具有了统一性,从而能设计适合的掩模来提取出缺陷部分,实现了对磁片表面刀纹缺陷的检测。实验结果表明,采用本文提出的方法可以实时完成对磁片表面刀纹缺陷的检测并且对于多种缺陷类型都适用。

关键词: 纹理特征, 表面刀纹缺陷检测, 图像处理, 实时

Abstract:

Surface inspection plays an important role in industrial production of product quality control, but the existing methods for the detection of magnetic material surface mark defects have their own limitations. The existing methods are too expensive and too slow to meet the needs of realtime detection in industrial production. In order to achieve the realtime and robust detection of surface mark defects, we propose a method in this paper based on the texture feature, which  makes the various defects have unity by enhancing the defective parts and inhibiting the background parts after image preprocessing. Then we detect surface mark defects by designing a suitable mask. The method proposed in this paper achieves the detection of surface mark defects. The experimental results show that it can be done in a realtime way and is useful for a variety of defects.

Key words: texture feature;detection of surface mark defects;image processing;realtime