近红外光谱技术在牛油生产过程中的应用

Application of near infrared technology in the butter production process

中文关键词:  近红外光谱法  牛油  支持向量机  偏最小二乘  水分

英文关键词:near infrared spectroscopy  butter  support vector machine  partial least squares  moisture

中文摘要:

      目的 利用便携式近红外(near infrared, NIR)光谱仪无损快速鉴别不同种类、不同部位的牛油原料及判别牛油熬制加工过程脱水程度。方法 利用NIR对不同种类、不同部位以及熬制过程不同时相点的牛油进行反射光谱采集, 通过光谱预处理、偏最小二乘法(partial least squares method, PLS)和支持向量机(support vector machine, SVM)探究鉴别牛油的不同种类和不同部位的预测模型和快速判别牛油熬制脱水程度。结果 利用PLS分析方法对不同种类、不同部位的牛油原料预测的正确率为75%; 利用SVM分析方法对牛油加工熬制过程中水分含量的预测正确率为78.6%。结论 使用便携式近红外光谱仪对不同种类、不同部位的牛油原料和牛油熬制加工过程中水分含量的无损检测方式在工业上作为大通量检测相关质控环节指标具有可行性。

英文摘要:

      Objective To achieve non-destructive and rapid identification of different types and different parts of butter raw materials by portable near infrared (NIR) spectrometer, in order to discriminate the degree of dehydration during processing of butter. Method The reflectance spectra of different types and different parts and the different time points of the tanning process were collected by NIR, and through spectral preprocessing, partial least squares method (PLS) and support vector machine (SVM) were used to explore predictive models for identifying different types and different parts of butter and to quickly determine the degree of dehydration of butter. Results The correct rate of prediction of different types and different parts of butter raw materials by PLS analysis method was 75%; and the correct rate of prediction of moisture content in the processing of butter processing by SVM analysis method was 78.6%. Conclusion The use of portable NIR spectrometer for non-destructive testing of moisture content in different types and different parts of butter raw materials and butter processing is feasible in the industry, which can be as a quality control link for large-throughput testing.