计算机工程与应用 ›› 2013, Vol. 49 ›› Issue (21): 24-28.

• 博士论坛 • 上一篇    下一篇

带钢表面缺陷图像的可拓分类算法

陈  跃1,2,张晓光1   

  1. 1.中国矿业大学 机电学院,江苏 徐州 221008
    2.徐州工程学院 机电学院,江苏 徐州 221008
  • 出版日期:2013-11-01 发布日期:2013-10-30

Classification of surface defect images of steel strip by extenics theory

CHEN Yue1,2, ZHANG Xiaoguang1   

  1. 1.School of Mechatronics Engineering, China University of Mining and Technology, Xuzhou, Jiangsu 221008, China
    2.School of Mechatronics Engineering, Xuzhou Institute of Technology, Xuzhou, Jiangsu 221008, China
  • Online:2013-11-01 Published:2013-10-30

摘要: 针对带钢表面缺陷难以识别和分类的问题,将可拓集合理论应用于带钢表面缺陷图像的分类,由缺陷图像的灰度共生矩阵计算出的能量、熵、惯性矩和相关性作为分类的特征向量,通过对大量图像数据的统计分析,确定不同缺陷类别的各参数经典域和节域,计算出待分类缺陷相对于各缺陷类别的加权关联度值,由最大值确定待分类缺陷所属类别。选择实际图像进行分类实验,仿真结果显示能够取得较好的分类效果。

关键词: 可拓理论, 缺陷图像分类, 特征参数选择

Abstract: In the light of difficulty in recognition and classification of steel strip surface defects, extenics theory is applied to classifying defects image of steel strip. Four parameters of defects image’s GLCM are calculated and selected as eigenvector. Then classical and extensional ranges of eigenvector are determined by statistical analysis of a number of data. Weight-added dependent degrees are calculated between unclassified defect image and each category of preclassified defects, and the maximum illuminates which category the unclassified defect image beongs to. Results of simulation declare the effectiveness by selecting unclassified steel strip surface defects image for experiments.

Key words: extenics theory, classification of defect images, selection of eigenvector