基于聚乙烯膜包装奶酪成分的NIRS检测研究

考察了聚乙烯包装膜对奶酪光谱的影响，提出了消除包装膜影响的新方案。探讨了近红外反射法直接检测带包装膜的奶酪成分的可行性。通过Norris导数滤波处理光谱，可以基本消除聚乙烯包装膜对奶酪光谱的影响，采用PLS结合MSC、求导等预处理建立了包装奶酪定量分析模型，其脂肪和蛋白质相关系数分别为0.928和0.952；建模标准差分别为0.240和0.355；预测标准差分别为0.326和0.219。与无包装奶酪和未滤波处理包装奶酪的模型比较，结果显示：与无包装奶酪模型差异极小，优于未采用Norris导数滤波处理的包装奶酪模型。实验表明近红外光谱分析技术可以在无损条件下快速检测包装奶酪中脂肪、蛋白质含量。

Study on Component Detection of Cheese Wrapped by Polyethylene Film Based on NIRS

The possibility of direct determination of fat and protein of wrapped cheese by near infrared spectroscopy was studied. The influences of polyethylene film on the spectra were discussed in order to detect the components of wrapped cheese. And the influences were eliminated using Norris derivation filter pretreatment means. The models for fat and protein of wrapped cheese were calibrated by partial least squares regression (PLS) following eliminating outline, spectral pretreatment, and PLS factors optimization. The best models gave standard errors for calibration of 0.240 and 0.355, standard errors for prediction of 0.326 and 0.219, and correlation coefficients of 0.928 and 0.952 for fat and protein of wrapped cheese, respectively. The results showed no difference from those by non-wrapped cheese’s models, and were better than wrapped cheese’s models without Norris derivation filter pretreatment. Based on the results, it was concluded that near infrared spectroscopy is a reliable, accurate and fast method for non-invasive measurement of wrapped cheese fat and protein.