核Fisher判别分析在棉麻纤维识别中的应用

Application of kernel Fisher discriminant analysis in cotton and bast fiber recognition

核Fisher判别分析方法的主要思想是把样本映射到某一特征空间,然后在此特征空间中进行Fisher线性判别.本文将Fisher判别分析方法用于棉、麻纤维的分类识别,实验结果表明:选择不同的核函数其判别效果不同,总体上核Fisher判别分析方法对棉麻纤维的识别率明显较高.

The main idea of the kernel Fisher discriminant analysis method is that the samples are mapped into a certain space,and then discriminated with Fisher linear discriminant in it.The kernel Fisher discriminant analysis method is used into classification and identification of cotton and bast fiber.The experimental results show that choices of different kernel function will lead to the different discriminant effect.On the whole,the recognition rate of kernel Fisher discriminant analysis method for cotton fiber and bast fiber is obviously higher.