摘要
通过图像处理技术把光学显微镜图像转化为谱线图。根据羊绒和羊毛谱线图呈现的区别,在谱线图上提取6个比对特征,通过比较6个特征的概率分布图以及对特征的相关性分析,最终用横向平均值和纵向离散度特征建立识别函数、识别羊绒和羊毛的混合样本。结果表明,此方法可以有效的识别羊绒和羊毛。
Digital image processing converts the digital images into spectral lines.According to the difference of the spectral line of wool and cashmere,six features are extracted from the spectral lines.With the comparison of the probability of the distributing graphs and the analysis of correlation of the six features,features of mean of the cross section and dispersion of the longitudinal direction are used to generate the discriminant function to identify wool and cashmere.Final result shows that it is an effective method in distinguishing cashmere from wool.
出处
《毛纺科技》
CAS
北大核心
2010年第5期38-41,共4页
Wool Textile Journal
关键词
羊毛
羊绒
谱线图
纹理图像
识别
wool
cashmere
spectral line
textural image
identification
