Study of the method of water-injected meat identifying based on low-field nuclear magnetic resonance

2017 3rd International Conference on Environmental Science and Material Application

生态环境科学;材料科学

Xu, Jianmei^1 ; Lin, Qing^2 ; Yang, Fang^1,2 ; Zheng, Zheng^3 ; Ai, Zhujun^3

Modern Technology Education Center, Hainan Medical University, Haikou

571199, China^1

College of Medical Information, Hainan Medical University, Haikou

571199, China^2

College of Tropical Medicine and Laboratory, Hainan Medical University, Haikou

571199, China^3

关键词: Cross validation;    Discriminant analysis methods;    Low field nuclear magnetic resonance;    Quantitative classifications;    Regular variations;    Spectral parameters;    T2 relaxation;    Transverse relaxation;   

Others  :  https://iopscience.iop.org/article/10.1088/1755-1315/108/4/042053/pdf
DOI  :  10.1088/1755-1315/108/4/042053

来源: IOP

【 摘 要 】

The aim of this study to apply low-field nuclear magnetic resonance technique was to study regular variation of the transverse relaxation spectral parameters of water-injected meat with the proportion of water injection. Based on this, the method of one-way ANOVA and discriminant analysis was used to analyse the differences between these parameters in the capacity of distinguishing water-injected proportion, and established a model for identifying water-injected meat. The results show that, except for T21b, T22eand T23b, the other parameters of the T2relaxation spectrum changed regularly with the change of water-injected proportion. The ability of different parameters to distinguish water-injected proportion was different. Based on S, P22and T23mas the prediction variable, the Fisher model and the Bayes model were established by discriminant analysis method, qualitative and quantitative classification of water-injected meat can be realized. The rate of correct discrimination of distinguished validation and cross validation were 88%, the model was stable.