计算机应用 ›› 2011, Vol. 31 ›› Issue (12): 3446-3448.

• 典型应用 • 上一篇    下一篇

基于多类支持向量机的棉花异性纤维分类方法

杨文柱,卢素魁,王思乐   

  1. 河北大学 数学与计算机学院,河北 保定 071002
  • 收稿日期:2011-05-10 修回日期:2011-06-17 发布日期:2011-12-12 出版日期:2011-12-01
  • 通讯作者: 杨文柱
  • 基金资助:
    国家自然科学基金;现代精细农业系统集成研究教育部重点实验室开放基金资助项目

Classification of cotton foreign fibers based on multi-class support vector machine

YANG Wen-zhu,LU Su-kui,WANG Si-le   

  1. College of Mathematics and Computer Science, Hebei University, Baoding Hebei 071002, China
  • Received:2011-05-10 Revised:2011-06-17 Online:2011-12-12 Published:2011-12-01
  • Contact: YANG Wen-zhu

摘要: 提出一种基于多类支持向量机的棉花异性纤维分类方法,以期解决棉花异性纤维的在线分类难题。该方法首先对异性纤维目标图像进行颜色、形状和纹理特征提取,形成用于精确描述异性纤维目标的特征向量;然后分别构建3种不同体系结构的多类支持向量机用于棉花异性纤维的分类;最后采用交叉验证法对所构建的3种多类支持向量机进行测试。测试结果表明,基于有向无环图的一对一多类支持向量机在分类精度和分类速度上更适合用于棉花异性纤维在线分类。

关键词: 异性纤维, 在线分类, 特征向量, 多类支持向量机, 留一交叉验证

Abstract: This paper proposed a new classification method based on Multi-class Support Vector Machine (MSVM) which aimed at solving the problems in online classification of cotton foreign fibers. Firstly the features of color, shape and texture of the foreign fiber objects were extracted to create the feature vectors. Secondly three kinds of multi-class support vector machines were constructed for foreign fiber classification. These three MSVMs were tested with the obtained feature vectors using leave-one-out cross validation. The experimental results show that the one-against-one directed acyclic graph MSVM is the fastest one and is fitter for online classification of foreign fibers.

Key words: foreign fiber, online classification, feature vector, multi-class support vector machine, leave-one-out cross validation