Methods of foreign fiber detecting based on PCA analyzing of infrared spectral images

Abstract

A set of near infrared high resolution spectral imaging system is set up, the infrared absorption properties of raw cotton and colorless foreign are analyzed through the system, and scheme of polypropylene fiber detection based on the near infrared spectral image is proposed; On this basis, reduce dimensions the spectral images through the principal component analysis, further improve the efficiency of colorless foreign detection. The experimental results show that the spectral images after reducing dimensions can be used to detect colorless or light color raw cotton fiber effectively.