摘要
旋切单板的纹理对缺陷的检测会产生干扰,本文提出一种改进的模糊C聚类均值(FCM)算法的旋切单板表面缺陷检测方法,该方法考虑了类内样本密度和类间距离作为综合参数,从而可以获得合理的初始聚类中心。该算法可以较好的检测出旋切单板表面纹理和缺陷信息。
This paper presents a modified Fuzzy C-Mean algorithm (FCM) for detection of texture and defects on the rotary-veneer surface. The sample density of the inter-class and the distances of intra-class are used in the system as comprehensive parameters to obtain the initial cluster centers.
出处
《自动化技术与应用》
2008年第11期79-82,85,共5页
Techniques of Automation and Applications
基金
江苏省教育厅自然科学基金资助项目(编号2005KJD520033)
关键词
旋切单板
模糊C均值聚类
缺陷检测
纹理
rotary-veneer
modified Fuzzy C-Mean algorithm (FCM)
defect detection
texture