Computer Engineering and Applications ›› 2009, Vol. 45 ›› Issue (23): 219-223.DOI: 10.3778/j.issn.1002-8331.2009.23.063

• 工程与应用 • Previous Articles     Next Articles

Automatic wood defects recognition based on texture extraction and support vector machine technology

ZHANG Zhao,YE Ning,YE Qiao-lin   

  1. School of Information Technology,Nanjing Forestry University,Nanjing 210037,China
  • Received:2008-04-25 Revised:2008-07-23 Online:2009-08-11 Published:2009-08-11
  • Contact: ZHANG Zhao

基于纹理提取和SVM技术的自动木材缺陷识别

张 召,业 宁,业巧林   

  1. 南京林业大学 信息科学技术学院,南京 210037
  • 通讯作者: 张 召

Abstract: Support Vector Machine(SVM) is a new method of pattern recognition,which has good generalization ability and outreach capacity.Automatic wood surface defects identification based on texture and support vector machine is studied.LBP texture feature extractor technology to complement down-dimensional processing of the wood image data is adopted,and the distribution traits of different wood surface defects is studied.With support vector machine classification algorithm,implementes fast and accurate automatic identification to judge whether the wood surface had any defects or not.And in locating the exact position of defects automatically is achieved.Many cross validation experiments indicted that support vector machine classification algorithm had strong recognition abilities to wood surface defects and the identification rates reached up to 96% upwards.

摘要: 要:支持向量机(SVM)是一种新的模式识别方法,有较好的泛化能力和推广能力。研究了基于纹理提取和支持向量机的自动木材表面缺陷的识别问题,借助LBP纹理特征提取技术实现对木材图像数据降维处理,并研究了木材表面不同类型缺陷的分布规律。利用支持向量机分类算法对木材表面有无缺陷进行了快速准确的自动识别,实现了木材表面缺陷的自动定位。多次交叉实验表明,SVM分类算法对木材表面缺陷具有较好的识别能力,识别率可达96%以上。

CLC Number: