摘要
对3个品种、3个部位的106个羊肉样品进行近红外光谱扫描,并测定其蛋白质、水分、脂肪含量,采用Unscrambler软件建立基于偏最小二乘法的近红外光谱预测模型。结果显示:样品水分含量近红外光谱校正决定系数为0.94,验证决定系数是0.86;蛋白质含量近红外光谱预测模型的校正决定系数为0.90,验证决定系数为0.72;脂肪含量近红外光谱校正决定系数0.81,验证决定系数0.64,由此可知近红外光谱用于羊肉品质检测具有可行性。本研究为羊肉化学成分的快速检测提供了基础。
106 mutton samples from three different parts of three breeds of sheep were collected,and near infrared spectroscopy predictive models were established.The content of protein,water and fat were determined and the models were got by partial least squares（PLS） of Unscrambler software.The results were as follows： the moisture’s R-square of calibration（R2C） and R-square of prediction（R2V） were 0.94 and 0.73;the protein’s R2C and R2V were 0.91 and 0.61;R2C of fat was 0.81,and R2V was 0.64,respectively.It was concluded that the near infrared spectroscopy could be used as a rapid tool to predict the mutton chemical composition,and the research provided basis for chemical determination of mutton.
出处
《核农学报》
CAS
CSCD
北大核心
2012年第3期500-504,共5页
Journal of Nuclear Agricultural Sciences
基金
公益性行业(农业)科研专项(200903043)
国家现代肉羊产业技术体系项目(CARS-39)
关键词
近红外光谱
偏最小二乘法
羊肉
化学成分
near infrared spectroscopy； partial least squares； mutton； chemical composition；
