Qualitative Identification And Quantitative Analysis Of Chicken Meat By Near Infrared Spectroscopy

Our country is the chicken meat production and consumption country, chicken production ranks second in the world, and chicken ranks behind pork as the second largest consumption of meat in china. Its traceability and edible quality measurement has been the focus of meat industry. The tradition detection methods of meat quality is chemical analysis, instrumental analysis and sensory analysis and so on, which need to consume a lot of manpowers, materials and financial resources. NIRS is different because of its characteristics of nondestructive, sensitive, rapid and low cost, which began to receive widespread attention in the meat industry. NIRS can not only realize the traditional chemical detection of meat composition, and can also realize qualitative identification of species, breeds and origins etc. This study applied NIRS to qualitative identification of chicken breeds, origins and quantitative analysis of characteristic components, and carried out the in-depth study on chemical metrology method of NIRS, in order to provide new methods and tools for rapid detection of chicken quality. The major conclusions are summarized as follows:(1) The qualitative identification of chicken from different breeds by NIRS was studied. The color, protein, fat, moisture and inosine acid content of chicken breast samples were determined, the near infrared spectrum was collected, and on this basis,the qualitative recognition effects of partial least squares discriminate model established by different sample states and different spectra pretreatment methods were compared and analyzed. The results showed that identification accuracy of different varieties of chicken breast can reach 90%, so it was feasible to quickly and nondestructively identify chicken from different breeds using NIRS.(2) The qualitative identification of chicken from different regions by NIRS was studied. The color, protein, fat and moisture content of chicken breast samples were determined, the near infrared spectrum was collected, and on this basis,the qualitative recognition effects of partial least squares discriminate model established by different sample states and different spectra pretreatment methods were compared and analyzed. The results showed that identification accuracy of different regions of chicken breast can reach 87.5% and 93.8%, so it was feasible to quickly and nondestructively identify chicken from different regions using NIRS.(3)The quantitative prediction of characteristic components in chicken meat by NIRS was studied. The inosine acid content and flavor amino acid content of chicken breast samples were determined, the near infrared spectrum was collected, and on this basis, the quantitative prediction effects of partial least squares discriminate model established by different sample states and different spectra pretreatment methods were compared and analyzed. The results showed that the quantitative analysis model of meat emulsion was superior to the quantitative analysis model of meat slice. The results also showed that the best spectral pretreatment method was Standard Normalized Variant(SNV) when quantitatively analyzed inosine acid content using NIRS. The correlation coefficient R2 c of calibration, correlation coefficient that it was feasible to predict inosine acid content using NIRS.