Shi Dongdong, Zhang Ping, Fan Zhiying, Tian Yuan, Liu Qingsheng
Feed Research Institute, Chinese Academy of Agricultural Sciences
The identification of the authenticity of protein materials and feed additive products is based on near-infrared fingerprint technology.
The research takes the study of qualified products as the core features. At first, we confirm and proof the authenticity of qualified product according to many reference methods. Then, choose pre-process method and select characteristic wavelength of the qualified spectra. Finally, through the spectra analysis of qualified products, we extract the fingerprint characters of their spectra to establish fingerprint identification model.
In order to guarantee the efficiency of identification, this study takes the unqualified products whose quality is very close to the qualified product as control. And the differences of Mahalanobis distance between the qualified sample and reference samples are the criteria of choosing the discriminate model. It can be seen from the research that NIR fingerprinting technique is rapid and efficient.